Papers

In multi-view multi-label learning (MVML), each instance is described by several heterogeneous feature representations and associated with multiple valid labels simultaneously. Although diverse MVML methods have been proposed over the last decade, most previous studies focus on leveraging the shared subspace across different views to represent the multi-view consensus information, while it is still an open issue whether such shared subspace representation is necessary when formulating the desired MVML model. In this paper, we propose a DeepGCN based View-Specific MVML method (D-VSM) which can bypass seeking for the shared subspace representation, and instead directly encoding the feature representation of each individual view through the deep GCN to couple with the information derived from the other views. Specifically, we first construct all instances under different feature representations into the corresponding feature graphs respectively, and then integrate them into a unified graph by integrating the different feature representations of each instance. Afterwards, the graph attention mechanism is adopted to aggregate and update all nodes on the unified graph to form structural representation for each instance, where both intra-view correlations and inter-view alignments have been jointly encoded to discover the underlying semantic relations. Finally, we derive a label confidence score for each instance by averaging the label confidence of its different feature representations with the multi-label soft margin loss. Extensive experiments have demonstrated that our proposed method significantly outperforms state-of-the-art methods.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Recent works have shown explainability and robustness are two crucial ingredients of trustworthy and reliable text classification. However, previous works usually address one of two aspects: i) how to extract accurate rationales for explainability while being beneficial to prediction; ii) how to make the predictive model robust to different types of adversarial attacks. Intuitively, a model that produces helpful explanations should be more robust against adversarial attacks, because we cannot trust the model that outputs explanations but changes its prediction under small perturbations. To this end, we propose a joint classification and rationale extraction model named AT-BMC. It includes two key mechanisms: mixed Adversarial Training (AT) is designed to use various perturbations in discrete and embedding space to improve the model’s robustness, and Boundary Match Constraint (BMC) helps to locate rationales more precisely with the guidance of boundary information. Performances on benchmark datasets demonstrate that the proposed AT-BMC outperforms baselines on both classification and rationale extraction by a large margin. Robustness analysis shows that the proposed AT-BMC decreases the attack success rate effectively by up to 69%. The results indicate that there are connections between robust models and better explanations.

poster session 3 @ red 4, poster session 12 @ red 4, oral session 3 @ red 4, poster session 3, poster session 12, oral session 3

Pre-trained Language Models (PLMs) have achieved great success in various Natural Language Processing (NLP) tasks under the pre-training and fine-tuning paradigm. With large quantities of parameters, PLMs are computation-intensive and resource-hungry. Hence, model pruning has been introduced to compress large-scale PLMs. However, most prior approaches only consider task-specific knowledge towards downstream tasks, but ignore the essential task-agnostic knowledge during pruning, which may cause catastrophic forgetting problem and lead to poor generalization ability. To maintain both task-agnostic and task-specific knowledge in our pruned model, we propose ContrAstive Pruning (CAP) under the paradigm of pre-training and fine-tuning. It is designed as a general framework, compatible with both structured and unstructured pruning. Unified in contrastive learn- ing, CAP enables the pruned model to learn from the pre-trained model for task-agnostic knowledge, and fine-tuned model for task-specific knowledge. Besides, to better retain the performance of the pruned model, the snapshots (i.e., the intermediate models at each pruning iteration) also serve as effective supervisions for pruning. Our extensive experiments show that adopting CAP consistently yields significant improvements, especially in extremely high sparsity scenarios. With only 3% model parameters reserved (i.e., 97% sparsity), CAP successfully achieves 99.2% and 96.3% of the original BERT performance in QQP and MNLI tasks. In addition, our probing experiments demonstrate that the model pruned by CAP tends to achieve better generalization ability.

poster session 3 @ red 4, poster session 12 @ red 4, oral session 3 @ red 4, poster session 3, poster session 12, oral session 3

Representation learning in dynamic graphs is a challenging problem because the topology of graph and node features vary at different time. This requires the model to be able to effectively capture both graph topology information and temporal information. Most existing works are built on recurrent neural networks (RNNs), which are used to exact temporal information of dynamic graphs, and thus they inherit the same drawbacks of RNNs. In this paper, we propose Learning to Evolve on Dynamic Graphs (LEDG) - a novel algorithm that jointly learns graph information and time information. Specifically, our approach utilizes gradient-based meta-learning to learn updating strategies that have better generalization ability than RNN on snapshots. It is model-agnostic and thus can train any message passing based graph neural network (GNN) on dynamic graphs. To enhance the representation power, we disentangle the embeddings into time embeddings and graph intrinsic embeddings. We conduct experiments on various datasets and down-stream tasks, and the experimental results validate the effectiveness of our method.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Pre-trained models have proved to be powerful in enhancing task-oriented dialog systems. However, current pre-training methods mainly focus on enhancing dialog understanding and generation tasks while neglecting the exploitation of dialog policy. In this paper, we propose GALAXY, a novel pre-trained dialog model that explicitly learns dialog policy from limited labeled dialogs and large-scale unlabeled dialog corpora via semi-supervised learning. Specifically, we introduce a dialog act prediction task for policy optimization during pre-training and employ a consistency regularization term to refine the learned representation with the help of unlabeled dialogs. We also implement a gating mechanism to weigh suitable unlabeled dialog samples. Empirical results show that GALAXY substantially improves the performance of task-oriented dialog systems, and achieves new state-of-the-art results on benchmark datasets: In-Car, MultiWOZ2.0 and MultiWOZ2.1, improving their end-to-end combined scores by 2.5, 5.3 and 5.5 points, respectively. We also show that GALAXY has a stronger few-shot ability than existing models under various low-resource settings. For reproducibility, we release the code and data at https://github.com/siat-nlp/GALAXY.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Acquiring high-quality temporal common sense (TCS) knowledge from free-form text is a crucial but challenging problem for event-centric natural language understanding, due to the language reporting bias problem: people rarely report the obvious events but highlight the special cases. For example, one may rarely report ``I get off the bed in 3 seconds'', but we can observe ``It takes me an hour to get off the bed every morning'' in text. This phenomenon can result in incomplete and biased TCS knowledge. Existing works address this issue mainly by exploiting the interactions among temporal dimensions (e.g., duration, the temporal relation between events) in a multi-task framework. However, this line of work suffers the limitation of implicit, inadequate, and unexplainable interactions. In this paper, we propose a novel Soft Logic Enhanced Event Temporal Reasoning (SLEER) model for acquiring unbiased TCS knowledge, in which the complementary relationship among dimensions are explicitly represented as logic rules and modeled by t-norm fuzzy logics. SLEER can utilize logic rules to regularize its inference process. Experimental results on two intrinsic datasets and two extrinsic datasets show the efficiency of our proposed method.

poster session 3 @ red 4, poster session 12 @ red 4, poster session 3, poster session 12

While text generated by current vision-language models may be accurate and syntactically correct, it is often general. Recent work has used optical character recognition to supplement visual information with text extracted from an image. In many cases, using text in the image improves the specificity and usefulness of generated text. We contend that vision-language models can benefit from additional information extracted from an image. We modify previous multimodal frameworks to accept relevant information from a number of auxiliary classifiers. In particular, we focus on person names as an additional set of tokens and create a novel image-caption dataset to facilitate captioning with person names. The dataset, Politicians and Athletes in Captions (PAC), consists of captioned images of well-known people in context. By fine-tuning pretrained models with this dataset, we demonstrate a model that can naturally integrate facial recognition tokens into generated text by training on limited data.

poster session 3 @ red 4, poster session 6 @ red 4, poster session 3, poster session 6

Generative models are used in a variety of applications that require diverse output. Yet, models are primarily optimised for sample fidelity and mode coverage. My work aims to increase the output diversity of generative models for multi-solution tasks. Previously, we analysed the use of generative models in artistic settings and how its objective diverges from distribution fitting. For specific use cases, we quantified the limitations of generative models. Future work will focus on adapting generative modelling for downstream tasks that require a diverse set of high-quality artefacts.

poster session 3 @ blue 4, poster session 6 @ blue 4, poster session 3, poster session 6

Recently, many approaches tackle the Unsupervised Domain Adaptive person re-identification (UDA re-ID) problem through pseudo-label-based contrastive learning. During training, a uni-centroid representation is obtained by simply averaging all the instance features from a cluster with the same pseudo label. However, a cluster may contain images with different identities (label noises) due to the imperfect clustering results, which makes the uni-centroid representation inappropriate. In this paper, we present a novel Multi-Centroid Memory (MCM) to adaptively capture different identity information within the cluster. MCM can effectively alleviate the issue of label noises by selecting proper positive/negative centroids for the query image. Moreover, we further propose two strategies to improve the contrastive learning process. First, we present a Domain-Specific Contrastive Learning (DSCL) mechanism to fully explore intra-domain information by comparing samples only from the same domain. Second, we propose Second-Order Nearest Interpolation (SONI) to obtain abundant and informative negative samples. We integrate MCM, DSCL, and SONI into a unified framework named Multi-Centroid Representation Network (MCRN). Extensive experiments demonstrate the superiority of MCRN over state-of-the-art approaches on multiple UDA re-ID tasks and fully unsupervised re-ID tasks.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

In this paper, we study the zero-shot sketch-based image retrieval (ZS-SBIR) task, which retrieves natural images related to sketch queries from unseen categories. In the literature, convolutional neural networks (CNNs) have become the de-facto standard and they are either trained end-to-end or used to extract pre-trained features for images and sketches. However, CNNs are limited in modeling the global structural information of objects due to the intrinsic locality of convolution operations. To this end, we propose a Transformer-based approach called \textit{\textbf{T}hree-Way \textbf{V}ision \textbf{T}ransformer (TVT)} to leverage the ability of Vision Transformer (ViT) to model global contexts due to the global self-attention mechanism. Going beyond simply applying ViT to this task, we propose a token-based strategy of adding fusion and distillation tokens and making them complementary to each other. Specifically, we integrate three ViTs, which are pre-trained on data of each modality, into a three-way pipeline through the processes of distillation and multi-modal hypersphere learning. The distillation process is proposed to supervise fusion ViT (ViT with an extra fusion token) with soft targets from modality-specific ViTs, which prevent fusion ViT from catastrophic forgetting. Furthermore, our method learns a multi-modal hypersphere by performing inter- and intra-modal alignment without loss of uniformity, which aims to bridge the modal gap between modalities of sketch and image and avoid the collapse in dimensions. Extensive experiments on three benchmark datasets, \ies, Sketchy, TU-Berlin, and QuickDraw, demonstrate the superiority of our TVT method over the state-of-the-art ZS-SBIR methods.

poster session 3 @ red 1, poster session 8 @ red 1, oral session 3 @ red 1, poster session 3, poster session 8, oral session 3

Point cloud has gained a lot of attention with the availability of a large amount of point cloud data and increasing applications like city planning and self-driving cars. However, current methods, often rely on labeled information and costly processing, such as converting point cloud to voxel. We propose a self-supervised learning approach to tackle these problems, combating labelling and additional memory cost issues. Our proposed method achieves results comparable to supervised and unsupervised baselines on the widely used benchmark datasets for self-supervised point cloud classification like ShapeNet, ModelNet10/40.

poster session 3 @ blue 5, poster session 7 @ blue 5, poster session 3, poster session 7

Occluded person re-identification is a challenging task as human body parts could be occluded by some obstacles (e.g. trees, cars, and pedestrians) in certain scenes. Some existing pose-guided methods solve this problem by aligning body parts according to graph matching, but these graph-based methods are not intuitive and complicated. Therefore, we propose a transformer-based Pose-guided Feature Disentangling (PFD) method by utilizing pose information to clearly disentangle semantic components (e.g. human body or joint parts) and selectively match non-occluded parts correspondingly. First, Vision Transformer (ViT) is used to extract the patch features with its strong capability. Second, to preliminarily disentangle the pose information from patch information, the matching and distributing mechanism is leveraged in Pose-guided Feature Aggregation (PFA) module. Third, a set of learnable semantic views are introduced in transformer decoder to implicitly enhance the disentangled body part features. However, those semantic views are not guaranteed to be related to the body without additional supervision. Therefore, Pose-View Matching (PVM) module is proposed to explicitly match visible body parts and automatically separate occlusion features. Fourth, to better prevent the interference of occlusions, we design a Pose-guided Push Loss to emphasize the features of visible body parts. Extensive experiments over five challenging datasets for two tasks (occluded and holistic Re-ID) demonstrate that our proposed PFD is superior promising, which performs favorably against state-of-the-art methods. Code is available at https://github.com/WangTaoAs/PFD_Net

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

Current state-of-the-art cross-lingual summarization models employ multi-task learning paradigm, which works on a shared vocabulary module and relies on the self-attention mechanism to attend among tokens in two languages. However, correlation learned by self-attention is often loose and implicit, inefficient in capturing crucial cross-lingual representations between languages. The matter worsens when performing on languages with separate morphological or structural features, making the cross-lingual alignment more challenging, resulting in the performance drop. To overcome this problem, we propose a novel Knowledge-Distillation-based framework for Cross-Lingual Summarization, seeking to explicitly construct cross-lingual correlation by distilling the knowledge of the monolingual summarization teacher into the cross-lingual summarization student. Since the representations of the teacher and the student lie on two different vector spaces, we further propose a Knowledge Distillation loss using Sinkhorn Divergence, an Optimal-Transport distance, to estimate the discrepancy between those teacher and student representations. Due to the intuitively geometric nature of Sinkhorn Divergence, the student model can productively learn to align its produced cross-lingual hidden states with monolingual hidden states, hence leading to a strong correlation between distant languages. Experiments on cross-lingual summarization datasets in pairs of distant languages demonstrate that our method outperforms state-of-the-art models under both high and low-resourced settings.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Translation-based approaches to planning allow for solving problems in complex and expressive formalisms via the means of highly efficient solvers for simpler formalisms. To be effective, these translations have to be constructed appropriately. The currently existing translation of the highly expressive formalism of HTN planning into the simplistic formalism of classical planning is not on par with the performance of current dedicated HTN planners. With our contributions in this paper, we close this gap: we describe new versions of the translation that reach the performance of state-of-the-art dedicated HTN planners. We present new translation techniques both for the special case of totally-ordered HTNs as well as for the general partially-ordered case. In the latter, we show hat our new translation is exponentially smaller than the previous one.

poster session 3 @ blue 4, poster session 12 @ blue 4, poster session 3, poster session 12

Despite their popularity for graph structured data, existing {\em Graph Neural Networks}~(GNNs) have inherent limitations for fundamental graph problems such as shortest paths, $k$-connectivity, minimum spanning tree and minimum cuts. In these instances, it is known that one needs GNNs of high depth, scaling at a polynomial rate with the number of nodes $n$, to provably encode the solution space, in turn affecting their statistical efficiency. In this work we propose a new hybrid architecture to overcome this limitation. Our proposed architecture that we call as {\GNNplus} networks involve a combination of multiple parallel low depth GNNs along with simple pooling layers involving low depth fully connected networks. We provably demonstrate that for many graph problems, the solution space can be encoded by {\GNNplus} networks using depth that scales only {\em poly-logarithmically} in the number of nodes. This also has statistical advantages that we demonstrate via generalization bounds for {\GNNplus} networks.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Federated learning (FL) and split learning (SL) are two popular distributed machine learning approaches. Both follow a model-to-data scenario; clients train and test machine learning models without sharing raw data. SL provides better model privacy than FL due to the machine learning model architecture split between clients and the server. Moreover, the split model makes SL a better option for resource-constrained environments. However, SL performs slower than FL due to the relay-based training across multiple clients. In this regard, this paper presents a novel approach, named splitfed learning (SFL), that amalgamates the two approaches eliminating their inherent drawbacks, along with a refined architectural configuration incorporating differential privacy and PixelDP to enhance data privacy and model robustness. Our analysis and empirical results demonstrate that (pure) SFL provides similar communication efficiency and test accuracy as SL while significantly decreasing its computation time per global epoch than in SL for multiple clients. Furthermore, as in SL, its communication efficiency over FL improves with the number of clients. Besides, the performance of SFL with privacy and robustness measures is further evaluated under extended experimental settings.

poster session 3 @ blue 3, poster session 8 @ blue 3, oral session 8 @ blue 3, poster session 3, poster session 8, oral session 8

After criminal recidivism or hiring machine learning mod-els have inflicted harm, participatory machine learning meth-ods are often used as a corrective positioning. However, lit-tle guidance exists on how to develop participatory machinelearning models throughout stages of the machine learningdevelopment life-cycle. Here we demonstrate how to co-design and partner with community groups, in the specificcase of feminicide data activism. We co-designed and piloteda machine learning model for the detection of media arti-cles about feminicide. This provides a feminist perspectiveon practicing participatory methods in a co-creation mind-set for the real-world scenario of monitoring violence againstwomen.

poster session 3 @ red 6, poster session 5 @ red 6, poster session 3, poster session 5

In point cloud generation and completion, previous methods for transforming latent features to point clouds are generally based on fully connected layers (FC-based) or folding operations (Folding-based). However, point clouds generated by FC-based methods are usually troubled by outliers and rough surfaces. For folding-based methods, their data flow is large, convergence speed is slow, and they are also hard to handle the generation of non-smooth surfaces. In this work, we propose AXform, an attention-based method to transform latent features to point clouds. AXform first generates points in an interim space, using a fully connected layer. These interim points are then aggregated to generate the target point cloud. AXform takes both parameter sharing and data flow into account, which makes it has fewer outliers, fewer network parameters, and a faster convergence speed. The points generated by AXform do not have the strong 2-manifold constraint, which improves the generation of non-smooth surfaces. When AXform is expanded to multiple branches for local generations, the centripetal constraint makes it has properties of self-clustering and space consistency, which further enables unsupervised semantic segmentation. We also adopt this scheme and design AXformNet for point cloud completion. Considerable experiments on different datasets show that our methods achieve state-of-the-art results.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

We address cross-species 3D face morphing (i.e., 3D face morphing from human to animal), a novel problem with promising applications in social media and movie industry. It remains challenging how to preserve target structural information and source fine-grained facial details simultaneously. To this end, we propose an Alignment-aware 3D Face Morphing (AFM) framework, which builds semantic-adaptive correspondence between source and target faces across species, via an alignment-aware controller mesh (Explicit Controller, EC) with explicit source/target mesh binding. Based on EC, we introduce Controller-Based Mapping (CBM), which builds semantic consistency between source and target faces according to the semantic importance of different face regions. Additionally, an inference-stage coarse-to-fine strategy is exploited to produce fine-grained meshes with rich facial details from rough meshes. Extensive experimental results in multiple people and animals demonstrate that our method produces high-quality deformation results.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Top-k sparsification method is popular and powerful forreducing the communication cost in Federated Learning(FL). However, according to our experimental observation, it spends most of the total communication cost on the index of the selected parameters (i.e., their position informa-tion), which is inefficient for FL training. To solve this problem, we propose a FL compression algorithm for convolution neural networks (CNNs), called SmartIdx, by extending the traditional Top-k largest variation selection strategy intothe convolution-kernel-based selection, to reduce the proportion of the index in the overall communication cost and thusachieve a high compression ratio. The basic idea of SmartIdx is to improve the 1:1 proportion relationship betweenthe value and index of the parameters to n:1, by regarding the convolution kernel as the basic selecting unit in parameter selection, which can potentially deliver more informationto the parameter server under the limited network traffic. Tothis end, a set of rules are designed for judging which kernel should be selected and the corresponding packaging strategies are also proposed for further improving the compressionratio. Experiments on mainstream CNNs and datasets showthat our proposed SmartIdx performs2.5×−69.2×higher compression ratio than the state-of-the-art FL compression algorithms without degrading model performance.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

Object detection in high-resolution satellite imagery is emerging as a scalable alternative to on-the-ground survey data collection in many environmental and socioeconomic monitoring applications. However, performing object detection over large geographies can still be prohibitively expensive due to the high cost of purchasing imagery and compute. Inspired by traditional survey data collection strategies, we propose an approach to estimate object count statistics over large geographies through sampling. Given a cost budget, our method selects a small number of representative areas by sampling from a learnable proposal distribution. Using importance sampling, we are able to accurately estimate object counts after processing only a small fraction of the images compared to an exhaustive approach. We show empirically that the proposed framework achieves strong performance on estimating the number of buildings in the United States and Africa, cars in Kenya, brick kilns in Bangladesh, and swimming pools in the U.S., while requiring as few as 0.01% of satellite images compared to an exhaustive approach.

poster session 3 @ red 6, poster session 12 @ red 6, oral session 12 @ red 6, poster session 3, poster session 12, oral session 12

As countries enter the endemic phase of COVID-19, people's risk of exposure to the virus is greater than ever. There is a need to make more informed decisions in our daily lives on avoiding crowded places. Crowd monitoring systems typically require costly infrastructure. We propose a crowd-sourced crowd monitoring platform which leverages user inputs to generate crowd counts and forecast location crowdedness. A key challenge for crowd-sourcing is a lack of incentive for users to contribute. We propose a Reinforcement Learning based dynamic incentive mechanism to optimally allocate rewards to encourage user participation.

poster session 3 @ blue 6, poster session 7 @ blue 6, poster session 3, poster session 7

Graph kernels are historically the most widely-used technique for graph classification tasks. However, these methods suffer from limited performance because of the hand-crafted combinatorial features of graphs. In recent years, graph neural networks (GNNs) have become the state-of-the-art method in downstream graph-related tasks due to their superior performance. Most GNNs are based on Message Passing Neural Network (MPNN) frameworks. However, recent studies show that MPNNs can not exceed the power of the Weisfeiler-Lehman (WL) algorithm in graph isomorphism test. To address the limitations of existing graph kernel and GNN methods, in this paper, we propose a novel GNN framework, termed Kernel Graph Neural Networks (KerGNNs), which integrates graph kernels into the message passing process of GNNs. Inspired by convolution filters in convolutional neural networks (CNNs), KerGNNs adopt trainable hidden graphs as graph filters which are combined with subgraphs to update node embeddings using graph kernels. In addition, we show that MPNNs can be viewed as special cases of KerGNNs. We apply KerGNNs to multiple graph-related tasks and use cross-validation to make fair comparisons with benchmarks. We show that our method achieves competitive performance compared with existing state-of-the-art methods, demonstrating the potential to increase the representation ability of GNNs. We also show that the trained graph filters in KerGNNs can reveal the local graph structures of the dataset, which significantly improves the model interpretability compared with conventional GNN models.

poster session 3 @ blue 2, poster session 8 @ blue 2, oral session 8 @ blue 2, poster session 3, poster session 8, oral session 8

Since no classical planner consistently outperforms all others, it is important to select a planner that works well for a given classical planning task. The two strongest approaches for planner selection use image and graph convolutional neural networks. They have the drawback that the learned models are complicated and uninterpretable. To obtain explainable models, we identify a small set of simple task features and show that elementary and interpretable machine learning techniques can use these features to solve approximately as many tasks as the complex approaches based on neural networks.

poster session 3 @ blue 4, poster session 12 @ blue 4, poster session 3, poster session 12

Given a query graph and a database of corpus graphs, a graph retrieval system aims to deliver the most relevant corpus graphs. Graph retrieval based on subgraph matching has a wide variety of applications, e.g., molecular fingerprint detection, circuit design, and question answering. In this setup, a corpus graph is relevant to a query graph, if this query graph is a subgraph of the corpus graph. Existing neural graph retrieval models compare the node or graph embeddings of the query-corpus pairs to compute the relevance scores between them. However, such models may not provide edge consistency between query corpus pairs. Moreover, they predominantly use symmetric relevance scores, which are not appropriate in the context of subgraph matching, since the underlying relevance score in subgraph search should be measured using the partial order induced by subgraph-supergraph relationship. Consequently, they show poor retrieval performance in the context of subgraph matching. In response, we propose ISONET, an interpretable neural edge alignment model, which is better able to learn the edge consistent mapping necessary for subgraph matching. Furthermore, we design a novel scoring mechanism which enforces an asymmetric relevance score specifically for subgraph matching. ISONET’s design enables it to identify the underlying subgraph in corpus graph, which is relevant to the given query graph. Our experiments on diverse datasets show that ISONET outperforms existing baseline models for graph retrieval. Additionally, ISONET can provide the interpretable alignments between query-corpus graph pairs during inference, despite being trained only using binary relevance labels during training, without any ground truth information about node or edge alignments.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Argumentation frameworks (AFs) are a core formalism in the field of formal argumentation. As most standard computational tasks regarding AFs are hard for the first or second level of the Polynomial Hierarchy, a variety of algorithmic approaches to achieve manageable runtimes have been considered in the past. Among them, the backdoor-approach and the treewidth-approach turned out to yield fixed-parameter tractable fragments. However, many applications yield high parameter values for these methods, often rendering them infeasible in practice. We introduce the backdoor-treewidth approach for abstract argumentation, combining the best of both worlds with a guaranteed parameter value that does not exceed the minimum of the backdoor- and treewidth-parameter. In particular, we formally define backdoor-treewidth and establish fixed-parameter tractability for standard reasoning tasks of abstract argumentation. Moreover, we provide systems to find and exploit backdoors of small width, and present the results of systematic experiments evaluating the new parameter.

poster session 3 @ blue 5, poster session 7 @ blue 5, oral session 7 @ blue 5, poster session 3, poster session 7, oral session 7

Federated learning (FL) is a popular technique to train machine learning (ML) models with decentralized data. Extensive works have studied the performance of the global model; however, it is still unclear how the training process affects the final test accuracy. Exacerbating this problem is the fact that FL executions differ significantly from traditional ML with heterogeneous data characteristics across clients, involving more hyperparameters. In this work, we show that the final test accuracy of FL is dramatically affected by the early phase of the training process, i.e., FL exhibits critical learning periods, in which small gradient errors can have irrecoverable impact on the final test accuracy. To further explain this phenomenon, we generalize the trace of the Fisher Information Matrix (FIM) to FL and define a new notation called FedFIM, a quantity reflecting the local curvature of each clients from the beginning of the training in FL. Our findings suggest that the initial learning phase plays a critical role in understanding the FL performance. This is in contrast to many existing works which generally do not connect the final accuracy of FL to the early phase training. Finally, seizing critical learning periods in FL is of independent interest and could be useful for other problems such as the choices of hyperparameters such as the number of client selected per round, batch size, and more, so as to improve the performance of FL training and testing.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

In this paper, we investigate the generalization problem of person re-identification (re-id), whose major challenge is the distribution shift on an unseen domain. As an important tool of regularizing the distribution, batch normalization (BN) has been widely used in existing methods. However, they neglect that BN is severely biased to the training domain and inevitably suffers the performance drop if directly generalized without being updated. To tackle this issue, we propose Batch Norm Test-time Adaption (BNTA), a novel re-id framework that applies the self-supervised strategy to update BN parameters adaptively. Specifically, BNTA quickly explores the domain-aware information within unlabeled target data before inference, and accordingly modulates the feature distribution normalized by BN to adapt to the target domain. This is accomplished by two designed self-supervised auxiliary tasks, namely part positioning and part nearest neighbor matching, which help the model mine the domain-aware information with respect to the structure and identity of body parts, respectively. To demonstrate the effectiveness of our method, we conduct extensive experiments on three re-id datasets and confirm the superior performance to the state-of-the-art methods.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

To accumulate knowledge and improve its policy of behaviour, a reinforcement learning agent can learn `off-policy' about policies that differ from the policy used to generate its experience. This is important to learn counterfactuals, or because the experience was generated out of its own control. However, off-policy learning is non-trivial, and standard reinforcement-learning algorithms can be unstable and divergent. In this paper we discuss a novel family of off-policy prediction algorithms which are convergent by construction. The idea is to first learn on-policy about the data-generating behaviour, and then bootstrap an off-policy value estimate on this on-policy estimate, thereby constructing a value estimate that is partially off-policy. This process can be repeated to build a chain of value functions, each time bootstrapping a new estimate on the previous estimate in the chain. Each step in the chain is stable and hence the complete algorithm is guaranteed to be stable. Under mild conditions this comes arbitrarily close to the off-policy TD solution when we increase the length of the chain. Hence it can compute the solution even in cases where off-policy TD diverges. We prove that the proposed scheme corresponds to an iterative decomposition of the inverse key matrix. Empirically we evaluate the idea on challenging MDPs such as Baird's counter example and observe favourable results.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

Reconfiguring two shortest paths in a graph means modifying one shortest path to the other by changing one vertex at a time, so that all the intermediate paths are also shortest paths. This problem has several natural applications, namely: (a) revamping road networks, (b) rerouting data packets in a synchronous multiprocessing setting, (c) the shipping container stowage problem, and (d) the train marshalling problem. When modelled as graph problems, (a) is the most general case while (b), (c) and (d) are restrictions to different graph classes. We show that (a) is intractable, even for relaxed variants of the problem. For (b), (c) and (d), we present efficient algorithms to solve the respective problems. We also generalise the problem to when at most k (for some k >= 2) contiguous vertices on a shortest path can be changed at a time.

poster session 3 @ blue 4, poster session 12 @ blue 4, poster session 3, poster session 12

Point cloud registration is a fundamental step for many tasks. In this paper, we propose a neural network named DetarNet to decouple the translation $t$ and rotation $R$, so as to overcome the performance degradation due to their mutual interference in point cloud registration. First, a Siamese Network based Progressive and Coherent Feature Drift (PCFD) module is proposed to align the source and target points in high-dimensional feature space, and accurately recover translation from the alignment process. Then we propose a Consensus Encoding Unit (CEU) to construct more distinguishable features for a set of putative correspondences. After that, a Spatial and Channel Attention (SCA) block is adopted to build a classification network for finding good correspondences. Finally, the rotation is obtained by Singular Value Decomposition (SVD). In this way, the proposed network decouples the estimation of translation and rotation, resulting in better performance for both of them. Experimental results demonstrate that the proposed DetarNet improves registration performance on both indoor and outdoor scenes. Our code will be available in \url{https://github.com/ZhiChen902/DetarNet}.

poster session 3 @ red 2, poster session 8 @ red 2, oral session 3 @ red 2, poster session 3, poster session 8, oral session 3

This paper presents RealTIS, a Real-time emergent Topic Identification System for user-generated content on the web via social networking services such as Twitter, Weibo, and Facebook. Without user intervention, our proposed RealTIS system can efficiently collect necessary social media posts, construct a quality topic summarization from the vast sea of data, and then automatically identify whether the emerging topics will be out-breaking or just fading into silence. RealTIS uses a time-sliding window to compute the statistics about the basic structure (motifs) variation of the propagation network for a specific topic. These statistics are then used to predict unusual shifts in correlations, make early warning and detect outbreak. Besides, this work also illustrates the mechanism by which our proposed system makes early warning happen.

poster session 3 @ blue 6, poster session 7 @ blue 6, poster session 3, poster session 7

In dialogue systems, utterances with similar semantics may have distinctive emotions under different contexts. Therefore, modeling long-range contextual emotional relationships with speaker dependency plays a crucial part in dialogue emotion recognition. Meanwhile, distinguishing the different emotion categories is non-trivial since they usually have semantically similar sentiments. To this end, we adopt supervised contrastive learning to make different emotions mutually exclusive to identify similar emotions better. Meanwhile, we utilize an auxiliary response generation task to enhance the model's ability of handling context information, thereby forcing the model to recognize emotions with similar semantics in diverse contexts. To achieve these objectives, we use the pre-trained encoder-decoder model BART as our backbone model since it is very suitable for both understanding and generation tasks. The experiments on four datasets demonstrate that our proposed model obtains significantly more favorable results than the state-of-the-art model in dialogue emotion recognition. The ablation study further demonstrates the effectiveness of supervised contrastive loss and generative loss.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Batch Normalization (BN) as an important component assists Deep Neural Networks achieving promising performance for extensive learning tasks by scaling distribution of feature representations within mini-batches. However, the application of BN suffers from performance degradation under the scenario of Unsupervised Domain Adaptation (UDA), since the estimated statistics fail to concurrently describe two different domains. In this paper, we develop a novel normalization technique, named Collaborative Normalization (CoN), for eliminating domain discrepancy and accelerating the model training of neural networks for UDA. Unlike typical strategies only exploiting domain-specific statistics during normalization, our CoN excavates cross-domain knowledge and simultaneously scales features from various domains by mimicking the merits of collaborative representation. Our CoN can be easily plugged into popular neural network backbones for cross-domain learning. On the one hand, \textbf{theoretical analysis} guarantees that models with CoN promote discriminability of feature representations and accelerate convergence rate; on the other hand, \textbf{empirical study} verifies that replacing BN with CoN in popular network backbones effectively improves classification accuracy in most learning tasks across three cross-domain visual benchmarks.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

Procedural text understanding requires machines to reason about entity states within the dynamical narratives. Current procedural text understanding approaches are commonly entity-wise, which separately track each entity and independently predict different states of each entity. Such an entity-wise paradigm does not consider the interaction between entities and their states. In this paper, we propose a new scene-wise paradigm for procedural text understanding, which jointly tracks states of all entities in a scene-by-scene manner. Based on this paradigm, we propose Scene Graph Reasoner (SGR), which introduces a series of dynamically evolving scene graphs to jointly formulate the evolution of entities, states and their associations throughout the narrative. In this way, the deep interactions between all entities and states can be jointly captured and simultaneously derived from scene graphs. Experiments show that SGR not only achieves the new state-of-the-art performance but also significantly accelerates the speed of reasoning.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Information cascade is typically formalized as a process of (simplified) discrete sequence of events, and recent approaches have tackled its prediction via variants of recurrent neural networks. However, the information diffusion process is essentially an evolving directed acyclic graph (DAG) in the continuous-time domain. In this paper, we propose a transformer enhanced Hawkes process (Hawkesformer), which links the hierarchical attention mechanism with Hawkes process to model the arrival stream of discrete events continuously. A two-level attention architecture is used to parameterize the intensity function of Hawkesformer, which captures the long-term dependencies between nodes in graph and better embeds the cascade evolution rate for modeling short-term outbreaks. Experimental results demonstrate the significant improvements of Hawkesformer over the state-of-the-art.

poster session 3 @ blue 5, poster session 7 @ blue 5, poster session 3, poster session 7

We introduce VAST, the Valence-Assessing Semantics Test, a novel intrinsic evaluation task for contextualized word embeddings (CWEs). Despite the widespread use of contextualizing language models (LMs), researchers have no intrinsic evaluation task for understanding the semantic quality of CWEs and their unique properties as related to contextualization, the change in the vector representation of a word based on surrounding words; tokenization, the breaking of uncommon words into subcomponents; and LM-specific geometry learned during training. VAST uses valence, the association of a word with pleasantness, to measure the correspondence of word-level LM semantics with widely used human judgments, and examines the effects of contextualization, tokenization, and LM-specific geometry. Because prior research has found that CWEs from OpenAI's 2019 English-language causal LM GPT-2 perform poorly on other intrinsic evaluations, we select GPT-2 as our primary subject, and include results showing that VAST is useful for 7 other LMs, and can be used in 7 languages. GPT-2 results show that the semantics of a word are more similar to the semantics of context in layers closer to model output, such that VAST scores diverge between our contextual settings, ranging from Pearson’s rho of .55 to .77 in layer 11. We also show that multiply tokenized words are not semantically encoded until layer 8, where they achieve Pearson’s rho of .46, indicating the presence of an encoding process for multiply tokenized words which differs from that of singly tokenized words, for which rho is highest in layer 0. We find that a few neurons with values having greater magnitude than the rest mask word-level semantics in GPT-2’s top layer, but that word-level semantics can be recovered by nullifying non-semantic principal components: Pearson’s rho in the top layer improves from .32 to .76. Downstream POS tagging and sentence classification experiments indicate that the GPT-2 uses these principal components for non-semantic purposes, such as to represent sentence-level syntax relevant to next-word prediction. After isolating semantics, we show the utility of VAST for understanding LM semantics via improvements over related work on four word similarity tasks, with a score of .50 on SimLex-999, better than the previous best of .45 for GPT-2. Finally, we show that 8 of 10 WEAT bias tests, which compare differences in word embedding associations between groups of words, exhibit more stereotype-congruent biases after isolating semantics, indicating that non-semantic structures in LMs also mask social biases.

poster session 3 @ red 4, poster session 12 @ red 4, poster session 3, poster session 12

Hyperspectral imaging is used for a wide range of tasks from medical diagnostics to crop monitoring, but traditional imagers are prohibitively expensive for widespread use. This research strives to democratize hyperspectral imaging by using machine learning to reconstruct hyperspectral volumes from snapshot imagers. I propose a tunable lens with varying amounts of defocus paired with 31-channel spectral filter array mounted on a CMOS camera. These images are then fed into a reconstruction network that aims to recover the full 31-channel hyperspectral volume from a few encoded images with different amounts of defocus.

poster session 3 @ red 6, poster session 5 @ red 6, poster session 3, poster session 5

Current researches on spoken language understanding (SLU) heavily are limited to a simple setting: the plain text-based SLU that takes the user utterance as input and generates its corresponding semantic frames (e.g., intent and slots). Unfortunately, such a simple setting may fail to work in complex real-world scenarios when an utterance is semantically ambiguous, which cannot be achieved by the text-based SLU models. In this paper, we first introduce a new and important task, Profile-based Spoken Language Understanding (ProSLU), which requires the model that not only relies on the plain text but also the supporting profile information to predict the correct intents and slots. To this end, we further introduce a large-scale human-annotated Chinese dataset with over 5K utterances and their corresponding supporting profile information (Knowledge Graph (KG), User Profile (UP), Context Awareness (CA)). In addition, we evaluate several state-of-the-art baseline models and explore a multi-level knowledge adapter to effectively incorporate profile information. Experimental results reveal that all existing text-based SLU models fail to work when the utterances are semantically ambiguous and our proposed framework can effectively fuse the supporting information for sentence-level intent detection and token-level slot filling. Finally, we summarize key challenges and provide new points for future directions, which hopes to facilitate the research.

poster session 3 @ red 5, poster session 8 @ red 5, oral session 8 @ red 5, poster session 3, poster session 8, oral session 8

In this work we continue to build upon recent advances in reinforcement learning for finite Markov processes. A common approach among previous existing algorithms, both single-actor and distributed, is to either clip rewards or to apply a transformation method on Q-functions to handle a large variety of magnitudes in real discounted returns. We theoretically show that one of the most successful methods may not yield an optimal policy if we have a non-deterministic process. As a solution, we argue that distributional reinforcement learning lends itself to remedy this situation completely. By the introduction of a conjugate distributional operator, we may handle a large class of transformations for real returns with guaranteed theoretical convergence. We propose an approximating single-actor algorithm based on this operator that trains agents directly on raw rewards using a proper distributional metric given by the Cramér distance. To evaluate its performance in a stochastic setting we train agents on a suite of 55 Atari 2600 games using sticky-actions and obtain state-of-the-art performance compared to other well-known algorithms in the Dopamine framework.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

We consider a novel queuing problem where the decision-maker must choose to accept or reject randomly arriving tasks into a no buffer queue which are processed by N identical servers. Each task has a price, which is a positive real number, and a class. Each class of task has a different price distribution, service rate, and arrives according to an inhomogenous Poisson process. The objective is to decide which tasks to accept so that the total price of tasks processed is maximised over a finite horizon. We formulate the problem using a discrete time Markov Decision Process (MDP) with a hybrid state space. We show that the optimal value function has a specific structure, which enables us to solve the hybrid MDP exactly. Moreover, we rigorously prove that as the gap between successive decision epochs grows smaller, the discrete time solution approaches the optimal solution to the original continuous time problem. To improve the scalability of our approach to a greater number of servers and task classes, we present an approximation based on state abstraction. We validate our approach on synthetic data, as well as a real financial fraud data set, which is the motivating application for this work.

poster session 3 @ blue 4, poster session 12 @ blue 4, poster session 3, poster session 12

Unpaired video-to-video translation aims to translate videos between a source and a target domain without the need of paired training data, making it more feasible for real applications. Unfortunately, the translated videos generally suffer from temporal and semantic inconsistency. To address this, many existing works adopt spatiotemporal consistency constraints incorporating temporal information based on motion estimation. However, the inaccuracies in the estimation of motion deteriorate the quality of the guidance towards spatiotemporal consistency, which leads to unstable translation. In this work, we propose a novel paradigm that regularizes the spatiotemporal consistency by synthesizing motions in input videos with the generated optical flow instead of estimating them. Therefore, the synthetic motion can be applied in the regularization paradigm to keep motions consistent across domains without the risk of errors in motion estimation. Thereafter, we utilize our unsupervised recycle and unsupervised spatial loss, guided by the pseudo-supervision provided by the synthetic optical flow, to accurately enforce spatiotemporal consistency in both domains. Experiments show that our method is versatile in various scenarios and achieves state-of-the-art performance in generating temporally and semantically consistent videos. Code is available at: https://github.com/wangkaihong/Unsup_Recycle_GAN/.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

Graph Convolutional Networks (GCNs) have been widely applied in various fields due to their significant power on processing graph-structured data. Typical GCN and its variants work under a homophily assumption (i.e., nodes with same class are prone to connect to each other), while ignoring the heterophily which exists in many real-world networks (i.e., nodes with different classes tend to form edges). Existing methods deal with heterophily by mainly aggregating higher-order neighborhoods or combing the immediate representations, which leads to noise and irrelevant information in the result. But these methods did not change the propagation mechanism which works under homophily assumption (that is a fundamental part of GCNs). This makes it difficult to distinguish the representation of nodes from different classes. To address this problem, in this paper we design a novel propagation mechanism, which can automatically change the propagation and aggregation process according to homophily or heterophily between node pairs. To adaptively learn the propagation process, we introduce two measurements of homophily degree between node pairs, which is learned based on topological and attribute information, respectively. Then we incorporate the learnable homophily degree into the graph convolution framework, which is trained in an end-to-end schema, enabling it to go beyond the assumption of homophily. More importantly, we theoretically prove that our model can constrain the similarity of representations between nodes according to their homophily degree. Experiments on seven real-world datasets demonstrate that this new approach outperforms the state-of-the-art methods under heterophily or low homophily, and gains competitive performance under homophily.

poster session 3 @ blue 2, poster session 8 @ blue 2, oral session 3 @ blue 2, poster session 3, poster session 8, oral session 3

Graph-structured datasets usually have irregular graph sizes and connectivities, rendering the use of recent data augmentation techniques, such as Mixup, difficult. To tackle this challenge, we present the first Mixup-like graph augmentation method called Graph Transplant, which mixes irregular graphs in data space. To be well defined on various scales of the graph, our method identifies the sub-structure as a mix unit that can preserve the local information. Since the mixup-based methods without special consideration of the context are prone to generate noisy samples, our method explicitly employs the node saliency information to select meaningful subgraphs and adaptively determine the labels. We extensively validate our method with diverse GNN architectures on multiple graph classification benchmark datasets from a wide range of graph domains of different sizes. Experimental results show the consistent superiority of our method over other basic data augmentation baselines. We also demonstrate that Graph Transplant enhances the performance in terms of robustness and model calibration.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Optimising the approximation of Average Precision (AP) has been widely studied for image retrieval. Limited by the definition of AP, such methods consider both negative and positive instances ranking before each positive instance. However, we claim that only penalizing negative instances before positive ones is enough, because the loss only comes from these negative instances. To this end, we propose a novel loss, namely Penalizing Negative instances before Positive ones (PNP), which can directly minimize the number of negative instances before each positive one. In addition, AP-based methods adopt a fixed and sub-optimal gradient assignment strategy. Therefore, we systematically investigate different gradient assignment solutions via constructing derivative functions of the loss, resulting in PNP-I with increasing derivative functions and PNP-D with decreasing ones. PNP-I focuses more on the hard positive instances by assigning larger gradients to them and tries to make all relevant instances closer. In contrast, PNP-D pays less attention to such instances and slowly corrects them. For most real-world data, one class usually contains several local clusters. PNP-I blindly gathers these clusters while PNP-D keeps them as they were. Therefore, PNP-D is more superior. Experiments on three standard retrieval datasets show consistent results with the above analysis. Extensive evaluations demonstrate that PNP-D achieves the state-of-the-art performance.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

As systems involving multiple agents are increasingly deployed, there is a growing need to diagnose failures in such systems. Model-Based Diagnosis (MBD) is a well-known AI technique to diagnose faults in systems. In this approach, a model of the diagnosed system is given, and the real system is observed. A failure is announced when the real system's output contradicts the model's expected output. The model is then used to deduce the defective components that explain the unexpected observation. MBD has been increasingly being deployed in distributed and multi-agent systems. In this survey, we summarize twenty years of research in the field of model-based diagnosis algorithms for MAS diagnosis. We depict three attributes that should be considered when examining MAS diagnosis: (1) The objective of the diagnosis. Either diagnosing faults in the MAS plans or diagnosing coordination faults. (2) Centralized vs. distributed. The diagnosis method could be applied either by a centralized agent or by the agents in a distributed manner. (3) Temporal vs. non-temporal. Temporal diagnosis is used to diagnose the MAS's temporal behaviors, whereas non-temporal diagnosis is used to diagnose the conduct based on a single observation. We survey diverse studies in MBD of MAS based on these attributes, and provide novel research challenges in this field for the AI community.

poster session 3 @ blue 6, poster session 7 @ blue 6, oral session 3 @ blue 6, poster session 3, poster session 7, oral session 3

Breast reconstruction surgery requires extensive planning, usually with a CT scan that helps surgeons identify which vessels are suitable for harvest. Currently, there is no quantitative method for preoperative planning. In this work, we successfully develop a Deep Learning algorithm to segment the vessels within the region of interest for breast reconstruction. Ultimately, this information will be used to determine the optimal reconstructive method (choice of vessels, extent of the free flap/harvested tissue) to reduce intra- and postoperative complication rates.

poster session 3 @ red 6, poster session 1 @ red 6, poster session 3, poster session 1

We study Policy-extended Value Function Approximator (PeVFA) in Reinforcement Learning (RL), which extends conventional value function approximator (VFA) to take as input not only the state (and action) but also an explicit policy representation. Such an extension enables PeVFA to preserve values of multiple policies at the same time and brings an appealing characteristic, i.e., \emph{value generalization among policies}. We formally analyze the value generalization under Generalized Policy Iteration (GPI). From theoretical and empirical lens, we show that generalized value estimates offered by PeVFA may have lower initial approximation error to true values of successive policies, which is expected to improve consecutive value approximation during GPI. Based on above clues, we introduce a new form of GPI with PeVFA which leverages the value generalization along policy improvement path. Moreover, we propose a representation learning framework for RL policy, providing several approaches to learn effective policy embeddings from policy network parameters or state-action pairs. In our experiments, we evaluate the efficacy of value generalization offered by PeVFA and policy representation learning in several OpenAI Gym continuous control tasks. For a representative instance of algorithm implementation, Proximal Policy Optimization (PPO) re-implemented under the paradigm of GPI with PeVFA achieves about 40\% performance improvement on its vanilla counterpart in most environments.

poster session 3 @ blue 1, poster session 7 @ blue 1, oral session 3 @ blue 1, poster session 3, poster session 7, oral session 3

Goal-oriented dialog policy learning algorithms aim to learn a dialog policy for selecting language actions based on the current dialog state. Deep reinforcement learning methods have been used for dialog policy learning. This work is motivated by the observation that, although dialog is a domain with rich contextual knowledge, reinforcement learning methods are ill-equipped to incorporate such knowledge into the dialog policy learning process. In this paper, we develop a deep reinforcement learning framework for goal-oriented dialog policy learning that learns user preferences from user goal data, while leveraging commonsense knowledge from people. The developed framework has been evaluated using a realistic dialog simulation platform. Compared with baselines from the literature and the ablations of our approach, we see significant improvements in learning efficiency and the quality of the computed action policies.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Sample efficiency is crucial for imitation learning methods to be applicable in real-world applications. Many studies improve sample efficiency by extending adversarial imitation to be off-policy regardless of the fact that these off-policy extensions could either change the original objective or involve complicated optimization. We revisit the foundation of adversarial imitation and propose an off-policy sample efficient approach that requires no adversarial training or min-max optimization. Our formulation capitalizes on two key insights: (1) the similarity between the Bellman equation and the stationary state-action distribution equation allows us to derive a novel temporal difference (TD) learning approach; and (2) the use of a deterministic policy simplifies the TD learning. Combined, these insights yield a practical algorithm, Deterministic and Discriminative Imitation (D2-Imitation), which oper- ates by first partitioning samples into two replay buffers and then learning a deterministic policy via off-policy reinforcement learning. Our empirical results show that D2-Imitation is effective in achieving good sample efficiency, outperforming several off-policy extension approaches of adversarial imitation on many control tasks.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

This study focuses on the Nesterov's accelerated quasi-Newton (NAQ) method in the context of deep neural networks (DNN) and its applications. The thesis objective is to confirm the robustness and efficiency of Nesterov's acceleration to quasi-Netwon (QN) methods by developing practical algorithms for different fields of optimization problems.

poster session 3 @ red 4, poster session 4 @ red 4, poster session 3, poster session 4

Recent research on graph neural network (GNN) models successfully applied GNNs to classical graph algorithms and combinatorial optimisation problems. This has numerous benefits, such as allowing applications of algorithms when preconditions are not satisfied, or reusing learned models when sufficient training data is not available or can't be generated. Unfortunately, a key hindrance of these approaches is their lack of explainability, since GNNs are black-box models that cannot be interpreted directly. In this work, we address this limitation by applying existing work on concept-based explanations to GNN models. We introduce \textit{concept-bottleneck GNNs}, which rely on a modification to the GNN readout mechanism. Using three case studies we demonstrate that: (i) our proposed model is capable of accurately learning concepts and extracting propositional formulas based on the learned concepts for each target class; (ii) our concept-based GNN models achieve comparative performance with state-of-the-art models; (iii) we can derive global graph concepts, without explicitly providing any supervision on graph-level concepts.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Single image generation (SIG), described as generating diverse samples that have the same visual content as the given natural image, is first introduced by SinGAN, which builds a pyramid of GANs to progressively learn the internal patch distribution of the single image. It shows excellent performance in a wide range of image manipulation tasks. However, SinGAN has some limitations. Firstly, due to lack of semantic information, SinGAN cannot handle the object images well as it does on the scene and texture images. Secondly, the independent progressive training scheme is time-consuming and easy to cause artifacts accumulation. To tackle these problems, in this paper, we dig into the single image generation problem and improve SinGAN by fully-utilization of internal and external priors. The main contributions of this paper include: 1) We interpret single image generation from the perspective of the general generative task, that is, to learn a diverse distribution from the Dirac distribution composed of a single image. In order to solve this non-trivial problem, we construct a regularized latent variable model to formulate SIG. To the best of our knowledge, it is the first time to give a clear formulation and optimization goal of SIG, and all the existing methods for SIG can be regarded as special cases of this model. 2) We design a novel Prior-based end-to-end training GAN (PetsGAN), which is infused with internal prior and external prior to overcome the problems of SinGAN. For one thing, we employ the pre-trained GAN model to inject external prior for image generation, which can alleviate the problem of lack of semantic information and generate natural, reasonable and diverse samples, even for the object image. For another, we fully-utilize the internal prior by a differential Patch Matching module and an effective reconstruction network to generate consistent and realistic texture. 3) We construct abundant of qualitative and quantitative experiments on three datasets. The experimental results show our method surpasses other methods on both generated image quality, diversity, and training speed. Moreover, we apply our method to other image manipulation tasks (e.g., style transfer, harmonization) and the results further prove the effectiveness and efficiency of our method.

poster session 3 @ red 1, poster session 8 @ red 1, oral session 8 @ red 1, poster session 3, poster session 8, oral session 8

Recently, deep learning techniques are soaring and have shown dramatic improvements in real-world noisy image denoising. However, the statistics of real noise generally vary with different camera sensors and in-camera signal processing pipelines. This will induce problems of most deep denoisers for the overfitting or degrading performance due to the noise discrepancy between the training and test sets. To remedy this issue, we propose a novel flexible and adaptive denoising network, coined as FADNet. Our FADNet is equipped with a plane dynamic filter module, which generates weight filters with flexibility that can adapt to the specific input and thereby impedes the FADNet from overfitting to the training data. Specifically, we exploit the advantage of the spatial and channel attention, and utilize this to devise a decoupling filter generation scheme. The generated filters are conditioned on the input and collaboratively applied to the decoded features for representation capability enhancement. We additionally introduce the Fourier transform and its inverse to guide the predicted weight filters to adapt to the noisy input with respect to the image contents. Experimental results demonstrate the superior denoising performances of the proposed FADNet versus the state-of-the-art. In contrast to the existing deep denoisers, our FADNet is not only flexible and efficient, but also exhibits a compelling generalization capability, enjoying tremendous potential for practical usage.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

Classical planning tasks are modelled in PDDL which is a schematic language based on first-order logic. Most of the current planners turn this lifted representation into a propositional one via a grounding process. However, grounding may cause an exponential blowup. Therefore it is important to investigate methods for searching for plans on the lifted level. To build a lifted state-based planner, it is necessary to invent lifted heuristics. We introduce maps between PDDL tasks preserving plans allowing to transform a PDDL task into a smaller one. We propose a novel method for computing lifted (admissible) delete-free relaxed heuristics via grounding of the smaller task and computing the (admissible) delete-free relaxed heuristics there. This allows us to transfer the knowledge about relaxed heuristics from the grounded level to the lifted level.

poster session 3 @ blue 4, poster session 12 @ blue 4, oral session 12 @ blue 4, poster session 3, poster session 12, oral session 12

Deep Neural Networks have memory and computational demands that often render them difficult to use in low-resource environments. Also, highly dense networks are over-parameterized and thus prone to overfitting. To address these problems, we introduce a novel algorithm that prunes (sparsifies) weights from the network by taking into account their magnitudes and gradients taken against a validation dataset. Unlike existing pruning methods, our method does not require the network model to be retrained once initial training is completed. On the CIFAR-10 dataset, our method reduced the number of paramters of MobileNet by a factor of 9X, from 14 million to 1.5 million, with just a 3.8% drop in accuracy.

poster session 3 @ blue 4, poster session 6 @ blue 4, poster session 3, poster session 6

My dissertation investigates the computation of Bayes-Nash equilibria in auctions via multiagent learning. A particular focus lies on the game-theoretic analysis of learned gradient dynamics in such markets. This requires overcoming several technical challenges like non-differentiable utility functions and infinite-dimensional strategy spaces. Positive results may open the door for wide-ranging applications in Market Design and the economic sciences.

poster session 3 @ blue 6, poster session 1 @ blue 6, poster session 3, poster session 1

Sound event detection aims to identify the sound events in the audio recordings, whose applications seem to be evident in our daily life, such as the surveillance and monitoring applications. In this paper, we present a novel framework for the detection task, by combining using several improvements. To compress the model efficiently while retaining the detection accuracy, the self-distillation paradigm is employed to improve offline training. To empower the machines with the ability of uncertainty estimation, the Monte Carlo dropout is used in our framework. Moreover, the inference data augmentation strategy is utilized to improve the robustness of the detection task. Lastly, we present an interactive interface, which can be used to visualize the detection and the uncertainty for the prediction. We hope our tool can be helpful for practical machine listening.

poster session 3 @ blue 6, poster session 7 @ blue 6, poster session 3, poster session 7

Every year in the US, millions of animals are run over by vehicles making wildlife vehicle collisions a real danger to both animals and human. In addition, road networks be-come abiotic barriers to wildlife migration between regions creating ripple effects on ecosystems. In this paper, a smart wildlife warning sign system (SWWS) is demonstrated, utilizing the technologies of Internet of Things, image recognition, data processing and visualization. This smart sign system is intended to prevent roadkill by warning drivers to slow down once sensors are triggered and simultaneously capture animal images via infrared cam-era. Data collection is conducted through local neural network model identification of wildlife images and saved along with metadata based on animal activity occurrence. Wildlife activity data can be exported wirelessly to cloud database to assist ecologists and government road agencies to investigate and analyze the wildlife activity and migration patterns over time.

poster session 3 @ blue 6, poster session 7 @ blue 6, poster session 3, poster session 7

The rapid development of drone technology is leading to more and more use cases being proposed. In response, regulators are drawing up drone traffic management frameworks. However, to design solutions that are efficient, fair, simple, non-manipulable, and scalable, we need market design and AI expertise. To this end, we introduce the drone traffic management problem as a new research challenge to the market design and AI communities. We present five design desiderata that we have derived from our interviews with stakeholders from the regulatory side as well as from public and private enterprises. Finally, we provide an overview of the solution space to point out possible directions for future research.

poster session 3 @ blue 6, poster session 7 @ blue 6, oral session 7 @ blue 6, poster session 3, poster session 7, oral session 7

The 2:4 structured sparsity pattern released by NVIDIA Ampere architecture, requiring four consecutive values containing at least two zeros, enables doubling math throughput for matrix multiplications. Recent works mainly focus on inference speedup via 2:4 sparsity while training acceleration has been largely overwhelmed where backpropagation consumes around 70% of the training time. However, unlike inference, training speedup with structured pruning is nontrivial due to the need to maintain the fidelity of gradients and reduce the additional overhead of performing 2:4 sparsity online. For the first time, this article proposes fully sparse training (FST) where `fully' indicates that ALL matrix multiplications in forward/backward propagation are structurally pruned while maintaining accuracy. To this end, we begin with saliency analysis, investigating the sensitivity of different sparse objects to structured pruning. Based on the observation of spatial similarity among activations, we propose pruning activations with fixed 2:4 masks. Moreover, an Information Restoration block is proposed to retrieve the lost information, which can be implemented by efficient gradient-shift operation. Evaluation of accuracy and efficiency shows that we can achieve 2× training acceleration with negligible accuracy degradation on challenging large-scale classification and detection tasks.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

In recent times, the social credit systems (SCS) and similar AI-driven mass surveillance systems have been deployed by the Chinese government in various regions. However, the discussions around the SCS are ambiguous: some people call them very controversial and a breach of human rights, while other people say that the SCS are very similar in structure to the company rankings or background checks on individuals in the United States. In reality, though, there is no monolith and there are different forms of SCS deployed in different regions of China. In this paper, I review the different models of the Chinese SCS. Then, I compare how the different systems are upholding or breaching China’s own AI Ethics guidelines.

poster session 3 @ blue 6, poster session 1 @ blue 6, poster session 3, poster session 1

The development of InSAR (satellite Interferometric Synthetic Aperture Radar) enables accurate monitoring of land surface deformations, and has led to advances of deformation forecast for preventing landslide, which is one of the severe geological disasters. Despite the unparalleled success, existing spatio-temporal models typically make predictions on static adjacency relationships, simplifying the conditional dependencies and neglecting the distributions of variables. To overcome those limitations, we propose a Distribution Aware Probabilistic Framework (DAPF), which learns manifold embeddings while maintaining the distribution of deformations. We obtain a dynamic adjacency matrix upon which we approximate the true posterior while emphasizing the spatio-temporal characteristics. Experimental results on real-world dataset validate the superior performance of our method.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

This research explores how to measure and predict student’s mental workload when using digital interactive textbooks. The current work focuses on analysis of clickstream data from the textbook in search of viewing patterns among students. It was found that students typically fit one of three viewing patterns. These patterns can be used in further research to inform creation of new interactive texts for improved student success.

poster session 3 @ blue 5, poster session 2 @ blue 5, poster session 3, poster session 2

So far, named entity recognition (NER) has been involved with three major types, including flat, overlapped (aka. nested), and discontinuous NER, which have mostly been studied individually. Recently, a growing interest has been built for unified NER, tackling the above three jobs concurrently with one single model. Current best-performing methods mainly include span-based and sequence-to-sequence models, where unfortunately the former merely focus on boundary identification and the latter may suffer from exposure bias. In this work, we present a novel alternative by modeling the unified NER as word-word relation classification, namely W^2NER. The architecture resolves the kernel bottleneck of unified NER by effectively modeling the neighboring relations between entity words with Next-Neighboring-Word (NNW) and Tail-Head-Word-* (THW-*) relations. Based on the W^2NER scheme we develop a neural framework, in which the unified NER is modeled as a 2D grid of word pairs. We then propose multi-granularity 2D convolutions for better refining the grid representations. Finally, a co-predictor is used to sufficiently reason the word-word relations. We perform extensive experiments on 14 widely-used benchmark datasets for flat, overlapped, and discontinuous NER (8 English and 6 Chinese datasets), where our model beats all the current top-performing baselines, pushing the state-of-the-art performances of unified NER.

poster session 3 @ red 4, poster session 12 @ red 4, poster session 3, poster session 12

Heuristics for lifted planning are not yet as informed as the best heuristics for ground planning. Recent work introduced the idea of using Datalog programs to compute the additive heuristic over lifted tasks. Based on this work, we show how to compute the more informed FF heuristic in a lifted manner. We extend the Datalog program with executable annotations that can also be used to define other delete-relaxation heuristics. In our experiments, we show that a planner using the lifted FF implementation produces state-of-the-art results for lifted planners. It also reduces the gap to state-of-the-art ground planners in domains where grounding is feasible.

poster session 3 @ blue 4, poster session 12 @ blue 4, poster session 3, poster session 12

My PhD research focus is to produce a critical review of literature in Algorithmic Impact Assessment (AIA) and to develop an AIA tool that can be used to evaluate potential unintended impact of AI systems.

poster session 3 @ blue 6, poster session 1 @ blue 6, poster session 3, poster session 1

Accurate electricity demand forecasts play a key role in sustainable power systems. To enable better decision-making especially for demand flexibility of the end-user, it is necessary to provide not only accurate but also understandable and actionable forecasts. To provide accurate forecasts Global Forecasting Models (GFM) that are trained across time series have shown superior results in many demand forecasting competitions and real-world applications recently, compared with univariate forecasting approaches. We aim to fill the gap between the accuracy and the interpretability in global forecasting approaches. In order to explain the global model forecasts, we propose Local Interpretable Model-agnostic Rule-based Explanations for Forecasting (LIMREF), which is a local explainer framework that produces k-optimal impact rules for a particular forecast, considering the global forecasting model as a black-box model, in a model-agnostic way. It provides different types of rules which explain the forecast of the global model and the counterfactual rules, which provide actionable insights for potential changes to obtain different outputs for given instances. We conduct experiments using a large-scale electricity demand dataset with exogenous features such as temperature and calendar effects. Here, we evaluate the quality of the explanations produced by the LIMREF framework in terms of both qualitative and quantitative aspects such as accuracy, fidelity and comprehensibility, and benchmark those against other local explainers.

poster session 3 @ red 6, poster session 10 @ red 6, poster session 3, poster session 10

Recent years have witnessed a flurry of research activity in graph matching, which aims at finding the correspondence of nodes across two graphs and lies at the heart of many artificial intelligence applications. However, matching heterogeneous graphs with partial overlap remains a challenging problem in real-world applications. This paper proposes the first practical learning-to-match method to meet this challenge. The proposed unsupervised method adopts a novel partial OT paradigm to learn a transport plan and node embeddings simultaneously. In a from-one-to-all manner, the entire learning procedure is decomposed into a series of easy-to-solve sub-procedures, each of which only handles the alignment of a single type of nodes. A mechanism for searching the transport mass is also proposed. Experimental results demonstrate that the proposed method outperforms state-of-the-art graph matching methods.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

State constraints in AI Planning globally restrict the legal environment states. Standard planning languages make closed-domain and closed-world assumptions. Here we address open-world state constraints formalized by planning over a description logic (DL) ontology. Previously, this combination of DL and planning has been investigated for the light-weight DL DL-Lite. Here we propose a novel compilation scheme into standard PDDL with derived predicates, which applies to more expressive DLs and is based on the rewritability of DL queries into Datalog with stratified negation. We also provide a new rewritability result for the DL Horn-ALCHOIQ, which allows us to apply our compilation scheme to quite expressive ontologies. In contrast, we show that in the slight extension Horn-SROIQ no such compilation is possible unless the weak exponential hierarchy collapses. Finally, we show that our approach can outperform previous work on existing benchmarks for planning with DL ontologies, and is feasible on new benchmarks taking advantage of more expressive ontologies.

poster session 3 @ blue 5, poster session 7 @ blue 5, oral session 3 @ blue 5, poster session 3, poster session 7, oral session 3

How does neural connectivity in autistic children differ from neural connectivity in healthy children or autistic youths? What patterns in global trade networks are shared across classes of goods, and how do these patterns change over time? Answering questions like these requires us to differentially describe groups of graphs: Given a set of graphs and a partition of these graphs into groups, discover what graphs in one group have in common, how they systematically differ from graphs in other groups, and how multiple groups of graphs are related. We refer to this task as graph group analysis, which seeks to describe similarities and differences between graph groups by means of statistically significant subgraphs. To perform graph group analysis, we introduce Gragra, which uses maximum entropy modeling to identify a non-redundant set of subgraphs with statistically significant associations to one or more graph groups. Through an extensive set of experiments on a wide range of synthetic and real-world graph groups, we confirm that Gragra works well in practice.

poster session 3 @ red 3, poster session 7 @ red 3, oral session 3 @ red 3, poster session 3, poster session 7, oral session 3

Human child trafficking has become a global epidemic with over 10 million children forced into labor or prostitution. In this paper, we propose the ChildrEN SafEty and Rescue (CENSER) system used by the Guria non-profit organization to retrieve trafficked children from brothels in India. The CENSER system is formed of the proposed Memory Augmented ScatterNet ResNet Hybrid (MSRHN) network trained on three databases containing images of trafficked children at different ages, their kins, and their sketches. The CENSER system encodes the input image of a child using the proposed Memory Augmented ScatterNet ResNet Hybrid (MSRHN) network and queries the encoding with the (i) Age, (ii) Kinship, and (iii) Sketch databases to establish the child's identity. The CENSER system can also predict if a child is a minor, which is used along with their identity to convince law enforcement to initiate the rescue operation. The MSRHN network is pre-trained on the KinFace database and then fine-tuned on the three databases. The performance of the proposed model is compared with several state-of-the-art methods.

poster session 3 @ red 6, poster session 10 @ red 6, poster session 3, poster session 10

Detecting 3D lanes from the camera is a rising problem for autonomous vehicles. In this task, the correct camera pose is the key to generating accurate lanes, which can transform an image from perspective-view to the top-view. With this transformation, we can get rid of the perspective effects so that 3D lanes would look similar and can accurately be fitted by low-order polynomials. However, mainstream 3D lane detectors rely on perfect camera poses provided by other sensors, which is expensive and encounters multi-sensor calibration issues. To overcome this problem, we propose to predict 3D lanes by estimating camera pose from a single image with a two-stage framework. The first stage aims at the camera pose task from perspective-view images. To improve pose estimation, we introduce an auxiliary 3D lane task and geometry constraints to benefit from multi-task learning, which enhances consistencies between 3D and 2D, as well as compatibility in the above two tasks. The second stage targets the 3D lane task. It uses previously estimated pose to generate top-view images containing distance-invariant lane appearances for predicting accurate 3D lanes. Experiments demonstrate that, without ground truth camera pose, our method outperforms the state-of-the-art perfect-camera-pose-based methods and has the fewest parameters and computations. Codes are available at https://github.com/liuruijin17/CLGo.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Despite the recent progress in Graph Neural Networks (GNNs), it remains challenging to explain the predictions made by GNNs. Existing explanation methods mainly focus on \emph{post-hoc} explanations where another explanatory model is employed to provide explanations for a trained GNN. The fact that \emph{post-hoc} methods fail to reveal the original reasoning process of GNNs raises the need of building GNNs with \emph{built-in} interpretability. In this work, we propose \textbf{Prot}otype \textbf{G}raph \textbf{N}eural \textbf{N}etwork (ProtGNN), which combines prototype learning with GNNs and provides a new perspective on the explanations of GNNs. In ProtGNN, the explanations are naturally derived from the case-based reasoning process and are actually used during classification. The prediction of ProtGNN is obtained by comparing the inputs to a few learned prototypes in the latent space. Furthermore, for better interpretability and higher efficiency, a novel conditional subgraph sampling module is incorporated to indicate which part of the input graph is most similar to each prototype in ProtGNN+. Finally, we evaluate our method on a wide range of datasets and perform concrete case studies. Extensive results show that ProtGNN and ProtGNN+ can provide inherent interpretability while achieving accuracy on par with the non-interpretable counterparts.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

3D object detection from point clouds data has become an indispensable part in autonomous driving. Previous works for processing point clouds lie in either projection or voxelization. However, projection-based methods suffer from information loss while voxelization-based methods bring huge computation. In this paper, we propose to encode point clouds into structured color image representation (SCIR) and utilize 2D CNN to fulfill the 3D detection task. Specifically, we use the structured color image encoding module to convert the irregular 3D point clouds into a squared 2D tensor image, where each point corresponds to a spatial point in the 3D space. Furthermore, in order to fit for the Euclidean structure, we apply feature normalization to parameterize the 2D tensor image onto a regular dense color image. Then, we conduct repeated multi-scale fusion with different levels so as to augment the initial features and learn scale-aware feature representations for box prediction. Extensive experiments on KITTI benchmark, Waymo Open Dataset and more challenging nuScenes dataset show that our proposed method yields decent results and demonstrate the effectiveness of such representations for point clouds.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Depth estimation is a crucial step for 3D reconstruction with panorama images in recent years. Panorama images maintain the complete spatial information but introduce distortion with equirectangular projection. In this paper, we propose an ACDNet based on the adaptively combined dilated convolution to predict the dense depth map for a monocular panoramic image. Specifically, we combine the convolution kernels with different dilations to extend the receptive field in the equirectangular projection. Meanwhile, we introduce an adaptive channel-wise fusion module to summarize the feature maps and get diverse attention areas in the receptive field along the channels. Due to the utilization of channel-wise attention in constructing the adaptive channel-wise fusion module, the network can capture and leverage the cross-channel contextual information efficiently. Finally, we conduct depth estimation experiments on three datasets (both virtual and real-world) and the experimental results demonstrate that our proposed ACDNet substantially outperforms the current state-of-the-art (SOTA) methods.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

As a popular deep neural networks (DNN) compression technique, knowledge distillation (KD) has attracted increasing attentions recently. Existing KD methods usually utilize one kind of knowledge in an intermediate layer of DNN for classification tasks to transfer useful information from cumbersome teacher networks to compact student networks. However, this paradigm is not very suitable for semantic segmentation, a comprehensive vision task based on both pixel-level and contextual information, since it cannot provide rich information for distillation. In this paper, we propose a novel multi-knowledge aggregation and transfer (MKAT) framework to comprehensively distill knowledge within an intermediate layer for semantic segmentation. Specifically, the proposed framework consists of three parts: Independent Transformers and Encoders module (ITE), Auxiliary Prediction Branch (APB), and Mutual Label Calibration (MLC) mechanism, which can take advantage of abundant knowledge from intermediate features. To demonstrate the effectiveness of our proposed approach, we conduct extensive experiments on three segmentation datasets: Pascal VOC, Cityscapes, and CamVid, showing that MKAT outperforms the other KD methods.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

Most approaches for goal recognition rely on specifications of the possible dynamics of the actor in the environment when pursuing a goal. These specifications suffer from two key issues. First, encoding these dynamics requires careful design by a domain expert, which is often not robust to noise at recognition time. Second, existing approaches often need costly real-time computations to reason about the likelihood of each potential goal. In this paper, we develop a framework that combines model-free reinforcement learning and goal recognition to alleviate the need for careful, manual domain design, and the need for costly online executions. This framework consists of two main stages: Offline learning of policies or utility functions for each potential goal, and online inference. We provide a first instance of this framework using tabular Q-learning for the learning stage, as well as three measures that can be used to perform the inference stage. The resulting instantiation achieves state-of-the-art performance against goal recognizers on standard evaluation domains and superior performance in noisy environments.

poster session 3 @ blue 4, poster session 12 @ blue 4, poster session 3, poster session 12

Generating new images with desired properties (e.g. new view/poses) from source images has been enthusiastically pursued recently, due to its wide range of potential applications. One way to ensure high-quality generation is to use multiple sources with complementary information such as different views of the same object. However, as source images are often misaligned due to the large disparities among the camera settings, strong assumptions have been made in the past with respect to the camera(s) or/and the object in interest, limiting the application of such techniques. Therefore, we propose a new general approach which models multiple types of variations among sources, such as view angles, poses, facial expressions, in a unified framework, so that it can be employed on datasets of vastly different nature. We verify our approach on a variety of data including humans bodies, faces, city scenes and 3D objects. Both the qualitative and quantitative results demonstrate the better performance of our method than the state of the art.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

In Apprenticeship Learning (AL), we are given a Markov Decision Process (MDP) without access to the cost function. Instead, we observe trajectories sampled by an expert that acts according to some policy. The goal is to find a policy that matches the expert's performance on some predefined set of cost functions. We introduce an online variant of AL (Online Apprenticeship Learning; OAL), where the agent is expected to perform comparably to the expert while interacting with the environment. We show that the OAL problem can be effectively solved by combining two mirror descent based no-regret algorithms: one for policy optimization and another for learning the worst case cost. By employing optimistic exploration, we derive a convergent algorithm with $O(\sqrt{K})$ regret, where $K$ is the number of interactions with the MDP, and an additional linear error term that depends on the amount of expert trajectories available. Importantly, our algorithm avoids the need to solve an MDP at each iteration, making it more practical compared to prior AL methods. Finally, we implement a deep variant of our algorithm which shares some similarities to GAIL \cite{ho2016generative}, but where the discriminator is replaced with the costs learned by the OAL problem. Our simulations suggest that OAL performs well in high dimensional control problems.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

Generating absent keyphrases, which do not appear in the input document, is challenging in the keyphrase prediction task. Most previous works treat the problem as an autoregressive sequence-to-sequence generation task, which demonstrates promising results for generating grammatically correct and fluent absent keyphrases. However, such an end-to-end process with a complete data-driven manner is unconstrained, which is prone to generate keyphrases inconsistent with the input document. In addition, the existing autoregressive decoding method makes the generation of keyphrases must be done from left to right, leading to slow speed during inference. In this paper, we propose a constrained absent keyphrase generation method in a prompt-based learning fashion. Specifically, the prompt will be created firstly based on the keywords, which are defined as the overlapping words between absent keyphrase and document. Then, a mask-predict decoder is used to complete the absent keyphrase on the constraint of prompt. Experiments on keyphrase generation benchmarks have demonstrated the effectiveness of our approach. In addition, we evaluate the performance of constrained absent keyphrases generation from an information retrieval perspective. The result shows that our approach can generate more consistent keyphrases, which can improve document retrieval performance. What’s more, with a non-autoregressive decoding manner, our model can speed up the absent keyphrase generation by 8.67× compared with the autoregressive method.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Estimating per-pixel motion between video frames, known as optical flow, is a long-standing problem in video understanding and analysis. Most contemporary optical flow techniques largely focus on addressing the cross-image matching with feature similarity, with few methods considering how to explicitly reason over the given scene for achieving a holistic motion understanding. In this work, taking a fresh perspective, we introduce a novel graph-based approach, called adaptive graph reasoning for optical flow (AGFlow), to emphasize the value of scene/context information in optical flow. Our key idea is to decouple the context reasoning from the matching procedure, and exploit scene information to effectively assist motion estimation by learning to reason over the adaptive graph. The proposed AGFlow can effectively exploit the context information and incorporate it within the matching procedure, producing more robust and accurate results. On both Sintel clean and final passes, our AGFlow achieves the best accuracy with EPE of 1.43 and 2.47 pixels, outperforming state-of-the-art approaches by 11.2% and 13.6%, respectively. Codes will be publicly available.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

The expression of displeasure on a consumer’s behalf towards an organization, product, or event is denoted via the speech act known as complaint. Customers typically post reviews on retail websites and various social media platforms about the products or services they purchase, and the reviews may include complaints about the products or services. Automatic detection of consumers' complaints about items or services they buy can be critical for organizations and online merchants. Since they can use this insight to meet the customers' requirements, including handling and addressing the complaints. Previous studies on Complaint Identification (CI) are limited to text. Images posted with the reviews can provide cues to better identify complaints, thus emphasizing the importance of incorporating multi-modal inputs into the process. Furthermore, the customer's emotional state has a significant impact on the complaint expression, since any speech act is generally influenced by emotions. As a result, the impact of emotion and sentiment on automatic complaint identification must also be investigated. One of the major contributions of this work is a new dataset- Complaint, Emotion, and Sentiment Annotated Multi-modal Amazon Reviews Dataset (CESAMARD) which is a collection of opinionated texts (reviews) and images of the products posted on the website of the retail giant Amazon. We present an attention-based multi-modal, adversarial multi-task deep neural network model for complaint detection to demonstrate the utility of the multi-modal dataset. Experimental results indicate that the multi-modality and multi-tasking complaint identification outperforms uni-modal and single-task variants.

poster session 3 @ red 6, poster session 12 @ red 6, poster session 3, poster session 12

Audio-driven one-shot talking face generation methods are usually trained on video resources of various persons. However, their created videos often suffer unnatural mouth shapes and asynchronous lips because those methods struggle to learn a consistent speech style from different speakers. We observe that it would be much easier to learn a consistent speech style from a specific speaker, which leads to authentic mouth movements. Hence, we propose a novel one-shot talking face generation framework by exploring consistent correlations between audio and visual motions from a specific speaker and then transferring audio-driven motion fields to a reference image. Specifically, we develop an Audio-Visual Correlation Transformer (AVCT) that aims to infer talking motions represented by keypoint based dense motion fields from an input audio. In particular, considering audio may come from different identities in deployment, we incorporate phonemes to represent audio signals. In this manner, our AVCT can inherently generalize to audio spoken by other identities. Moreover, as face keypoints are used to represent speakers, AVCT is agnostic against appearances of the training speaker, and thus allows us to manipulate face images of different identities readily. Considering different face shapes lead to different motions, a motion field transfer module is exploited to reduce the audio-driven dense motion field gap between the training identity and the one-shot reference. Once we obtained the dense motion field of the reference image, we employ an image renderer to generate its talking face videos from an audio clip. Thanks to our learned consistent speaking style, our method generates authentic mouth shapes and vivid movements. Extensive experiments demonstrate that our synthesized videos outperform the state-of-the-art in terms of visual quality and lip-sync.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

Optimizing the combustion efficiency of a thermal power generating unit (TPGU) is a highly challenging and critical task in the energy industry. We develop a new data-driven AI system, namely DeepThermal, to optimize the combustion control strategy for TPGUs. At its core, is a new model-based offline reinforcement learning (RL) framework, called MORE, which leverages historical operational data of a TGPU to solve a highly complex constrained Markov decision process problem via purely offline training. In DeepThermal, we first learn a data-driven combustion process simulator from the offline dataset. The RL agent of MORE is then trained by combining real historical data as well as carefully filtered and processed simulation data through a novel restrictive exploration scheme. DeepThermal has been successfully deployed in four large coal-fired thermal power plants in China. Real-world experiments show that DeepThermal effectively improves the combustion efficiency of TPGUs. We also report the superior performance of MORE by comparing with the state-of-the-art algorithms on the standard offline RL benchmarks.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

We propose a theoretical framework of multi-way similarity to model real-valued data into hypergraphs for clustering via spectral embedding. For graph cut based spectral clustering, it is common to model real-valued data into graph by modeling pairwise similarities using kernel function. This is because the kernel function has a theoretical connection to the graph cut. For problems where using multi-way similarities are more suitable than pairwise ones, it is natural to model as a hypergraph, which is generalization of a graph. However, although the hypergraph cut is well-studied, there is not yet established a hypergraph cut based framework to model multi-way similarity. In this paper, we formulate multi-way similarities by exploiting the theoretical foundation of kernel function. We show a theoretical connection between our formulation and hypergraph cut in two ways, generalizing both weighted kernel $k$-means and the heat kernel, by which we justify our formulation. We also provide a fast algorithm for spectral clustering. Our algorithm empirically shows better performance than existing graph and other heuristic modeling methods.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Binary optimization plays an important role in mathematical optimization and has various applications in computer vision and machine learning. Generally speaking, binary optimization problems are NP-hard and difficult to solve due to the binary constraints, especially when the number of variables is very large. Existing methods often suffer from high computational costs or large accumulated quantization errors, or are only designed for specific tasks. In this paper, we propose an efficient algorithm, named Discrete Principal Coordinate Descent (DPCD), to find effective approximate solutions for general binary optimization problems. The proposed algorithm supports a wide range of empirical objective functions with/without restrictions on the numbers of 1s and -1s in the binary variables. Furthermore, the theoretical convergence is proven, and the explicit convergence rates are derived for objective functions with Lipschitz continuous gradients, which are commonly adopted in practice. Extensive experiments on several tasks and large-scale datasets demonstrate the superiority of the proposed algorithm over several state-of-the-art methods in terms of both effectiveness and efficiency.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

Reward-Weighted Regression (RWR) belongs to a family of widely known iterative Reinforcement Learning algorithms based on the Expectation-Maximization framework. In this family, learning at each iteration consists of sampling a batch of trajectories using the current policy and fitting a new policy to maximize a return-weighted log-likelihood of actions. Although RWR is known to yield monotonic improvement of the policy under certain circumstances, whether and under which conditions RWR converges to the optimal policy have remained open questions. In this paper, we provide for the first time a proof that RWR converges to a global optimum when no function approximation is used, in a general compact setting. Furthermore, for the simpler case with finite state and action spaces we prove R-linear convergence of the state-value function to the optimum.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

Over the last few years, YouTube Kids has emerged as one of the highly competitive alternatives to television for children's entertainment. Consequently, YouTube Kids' content should receive an additional level of scrutiny to ensure children's safety. While research on detecting offensive or inappropriate content for kids is gaining momentum, little or no current work exists that investigates to what extent AI applications can (accidentally) introduce content that is inappropriate for kids. In this paper, we present a novel (and troubling) finding that well-known automatic speech recognition (ASR) systems may produce text content highly inappropriate for kids while transcribing YouTube Kids' videos. We dub this phenomenon as inappropriate content hallucinations. Our analyses suggest that such hallucinations are far from occasional, and the ASR systems often produce them with high confidence. We release a first-of-its-kind data set of audios for which the existing state-of-the-art ASR systems hallucinate inappropriate content for kids. In addition, we demonstrate that some of these errors can be fixed using language models.

poster session 3 @ red 6, poster session 10 @ red 6, poster session 3, poster session 10

Neural network quantization has become increasingly popular due to efficient memory consumption and faster computation resulting from bitwise operations on the quantized networks. Even though they exhibit excellent generalization capabilities, their robustness properties are not well-understood. In this work, we systematically study the robustness of quantized networks against gradient based adversarial attacks and demonstrate that these quantized models suffer from gradient vanishing issues and show a fake sense of robustness. By attributing gradient vanishing to poor forward-backward signal propagation in the trained network, we introduce a simple temperature scaling approach to mitigate this issue while preserving the decision boundary. Despite being a simple modification to existing gradient based adversarial attacks, experiments on multiple image classification datasets with multiple network architectures demonstrate that our temperature scaled attacks obtain near-perfect success rate on quantized networks while outperforming original attacks on adversarially trained models as well as floating-point networks.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

This paper studies a weakly supervised domain adaptation (WSDA) problem, where we only have access to the source domain with noisy labels, from which we need to transfer useful information to the unlabeled target domain. Although there have been a few studies on this problem, most of them only exploit unidirectional relationships from the source domain to the target domain. In this paper, we propose a universal paradigm called GearNet to exploit bilateral relationships between the two domains. Specifically, we take the two domains as different inputs to train two models alternately, and a symmetrical Kullback-Leibler loss is used for selectively matching the predictions of the two models in the same domain. This interactive learning schema enables implicit label noise canceling and exploit correlations between the source and target domains. Therefore, our GearNet has the great potential to boost the performance of a wide range of existing WSDA methods. Comprehensive experimental results show that the performance of existing methods can be significantly improved by equipping with our GearNet.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

This demo paper presents a design and implementation of a system AnomalyKiTS for detecting anomalies from time series data for the purpose of offering a broad range of algorithms to the end user, with special focus on unsupervised/semi-supervised learning. Given an input time series, AnomalyKiTS provides four categories of model building capabilities followed by an enrichment module that helps to label anomaly. AnomalyKiTS also supports a wide range of execution engines to meet the diverse need of anomaly workloads such as Serveless for CPU intensive work, GPU for deep-learning model training, etc.

poster session 3 @ blue 6, poster session 7 @ blue 6, poster session 3, poster session 7

Capturing visual similarity among images is the core of many computer vision and pattern recognition tasks. This problem can be formulated in such a paradigm called metric learning. Most research in the area has been mainly focusing on improving the loss functions and similarity measures. However, due to the ignoring of geometric structure, existing methods often lead to sub-optimal results. Thus, several recent research methods took advantage of Wasserstein distance between batches of samples to characterize the spacial geometry. Although these approaches can achieve enhanced performance, the aggregation over batches definitely hinders Wasserstein distance's superior measure capability and leads to high computational complexity. To address this limitation, we propose a novel Deep Wasserstein Metric Learning framework, which employs Wasserstein distance to precisely capture the relationship among various images under ranking-based loss functions such as contrastive loss and triplet loss. Our method directly computes the distance between images, considering the geometry at a finer granularity than batch level. Furthermore, we introduce a new efficient algorithm using Sinkhorn approximation and Wasserstein measure coreset. The experimental results demonstrate the improvements of our framework over various baselines in different applications and benchmark datasets.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

The field of Temporal Planning (TP) is receiving increasing interest for its many real-world applications. Most of the literature focuses on the TP problem of finding a plan, with algorithms that are not guaranteed to terminate when the problem admits no solution. In this paper, we present sound and complete decision procedures that address the dual problem of proving that no plan exists, which has important applications in oversubscription, model validation and optimization. We focus on the expressive and practically relevant semantics of action non-self-overlapping, recently proved to be PSPACE-complete. For this subclass, we propose two approaches: a reduction of the planning problem to model-checking of Timed Transition Systems, and a heuristic-search algorithm where temporal constraints are represented by Difference Bound Matrices. We implemented the approaches, and carried out an experimental evaluation against other state-of-the-art TP tools. On benchmarks that admit no plans, both approaches dramatically outperform the other planners, while the heuristic-search algorithm remains competitive on solvable benchmarks.

poster session 3 @ blue 4, poster session 12 @ blue 4, oral session 3 @ blue 4, poster session 3, poster session 12, oral session 3

Deep representation learning has succeeded in several fields. However, pre-trained deep representations are usually biased and make downstream models sensitive to different attributes. In this work, we propose a post-processing unsupervised deep representation debiasing algorithm, DeepMinMax, which can obtain unbiased representations directly from pre-trained representations without re-training or fine-tuning the entire model. The experimental results on synthetic and real-world datasets indicate that DeepMinMax outperforms the existing state-of-the-art algorithms on downstream tasks.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

Our research aims to reduce the cost of pressure sensor calibration through machine learning. Pressure sensor calibration is a standard process whereby freshly manufactured pressure sensors are subjected to various controlled temperature and pressure setpoints to compute a mapping between the sensor's output and true pressure. Traditionally this mapping is calculated by fitting a polynomial with calibration data. Obtaining this data is costly since a large spectrum of temperature and pressure setpoints are required to model the sensor's behavior. We present a machine learning approach to predict a pre-defined calibration polynomial's parameters while requiring only one-third of the calibration data. Our method learns a pattern from past calibration sessions to predict the calibration polynomial's parameters from partial calibration setpoints for any newly manufactured sensor. We design a novel polynomial hypernetwork coupled with Fourier features and a weighted loss to solve this problem. We perform extensive evaluations and show that the current industry-standard method fails under similar conditions. In contrast, our approach saves two-thirds of the calibration time and cost. Furthermore, we conduct comprehensive ablations to study the effect of Fourier mapping and weighted loss. Code and a novel calibration dataset validated by calibration engineers are also made public.

poster session 3 @ red 6, poster session 10 @ red 6, poster session 3, poster session 10

Existing graph neural networks (GNNs) largely rely on node embeddings, which represent a node as a vector by its identity, type, or content. However, graphs with unattributed nodes widely exist in real-world applications (e.g., anonymized social networks). Previous GNNs either assign random labels to nodes (which introduces artefacts to the GNN) or assign one embedding to all nodes (which fails to explicitly distinguish one node from another). Further, when these GNNs are applied to unattributed node classification problems, they have an undesired equivariance property, which are fundamentally unable to address the data with multiple possible outputs. In this paper, we analyze the limitation of existing approaches to node classification problems. Inspired by our analysis, we propose a generalized equivariance property and a Preferential Labeling technique that satisfies the desired property asymptotically. Experimental results show that we achieve high performance in several unattributed node classification tasks.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Key information extraction (KIE) from document images requires understanding the contextual and spatial semantics of texts in two-dimensional (2D) space. Many recent studies try to solve the task by developing pre-trained language models focusing on combining visual features from document images with texts and their layout. On the other hand, this paper tackles the problem by going back to the basic: effective combination of text and layout. Specifically, we propose a pre-trained language model, named BROS (BERT Relying On Spatiality), that encodes relative positions of texts in 2D space and learns from unlabeled documents with area-masking strategy. With this optimized training scheme for understanding texts in 2D space, BROS shows comparable or better performance compared to previous methods on four KIE benchmarks (FUNSD, SROIE*, CORD, and SciTSR) without relying on visual features. This paper also reveals two real-world challenges in KIE tasks-(1) minimizing the error from incorrect text ordering and (2) efficient learning from fewer downstream examples-and demonstrates the superiority of BROS over previous methods. Code will be available at https://github.com/clovaai.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Traffic forecasting is one of the most popular spatio-temporal tasks in the field of machine learning. A prevalent approach in the field is to combine graph convolutional networks and recurrent neural networks for the spatio-temporal processing. There has been fierce competition and many novel methods have been proposed. In this paper, we present the method of spatio-temporal graph neural controlled differential equation (STG-NCDE). Neural controlled differential equations (NCDEs) are a breakthrough concept for processing sequential data. We extend the concept and design two NCDEs: one for the temporal processing and the other for the spatial processing. After that, we combine them into a single framework. We conduct experiments with 6 benchmark datasets and 20 baselines. STG-NCDE shows the best accuracy in all cases, outperforming all those 20 baselines by non-trivial margins.

poster session 3 @ blue 2, poster session 8 @ blue 2, oral session 8 @ blue 2, poster session 3, poster session 8, oral session 8

Inspired by the recent success of self-supervised methods applied on images, self-supervised learning on graph structured data has seen rapid growth especially centered on augmentation-based contrastive methods. However, we argue that without carefully designed augmentation techniques, augmentations on graphs may behave arbitrarily in that the underlying semantics of graphs can drastically change. As a consequence, the performance of existing augmentation-based methods is highly dependent on the choice of augmentation scheme, i.e., augmentation hyperparameters and combinations of augmentation. In this paper, we propose a novel augmentation-free self-supervised learning framework for graphs, named AFGRL. Specifically, we generate an alternative view of a graph by discovering nodes that share the local structural information and the global semantics with the graph. Extensive experiments towards various node-level tasks, i.e., node classification, clustering, and similarity search on various real-world datasets demonstrate the superiority of AFGRL. The source code for AFGRL is available at https://github.com/Namkyeong/AFGRL.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Existing permutation-invariant methods can be divided into two categories according to the aggregation scope, i.e. global aggregation and local one. Although the global aggregation methods, e. g., PointNet and Deep Sets, get involved in simpler structures, their performance is poorer than the local aggregation ones like PointNet++ and Point Transformer. It remains an open problem whether there exists a global aggregation method with a simple structure, competitive performance, and even much fewer parameters. In this paper, we propose a novel global aggregation permutation-invariant network based on dual MLP dot-product, called DuMLP-Pin, which is capable of being employed to extract features for set inputs, including unordered or unstructured pixel, attribute, and point cloud data sets. We strictly prove that any permutation-invariant function implemented by DuMLP-Pin can be decomposed into two or more permutation-equivariant ones in a dot-product way as the cardinality of the given input set is greater than a threshold. We also show that the DuMLP-Pin can be viewed as Deep Sets with strong constraints under certain conditions. The performance of DuMLP-Pin is evaluated on several different tasks with diverse data sets. The experimental results demonstrate that our DuMLP-Pin achieves the best results on the two classification problems for pixel sets and attribute sets. On both the point cloud classification and the part segmentation, the accuracy of DuMLP-Pin is very close to the so-far best-performing local aggregation method with only a 1-2% difference, while the number of required parameters is significantly reduced by more than 85% in classification and 69% in segmentation, respectively.

poster session 3 @ red 2, poster session 8 @ red 2, oral session 8 @ red 2, poster session 3, poster session 8, oral session 8

Graph neural networks (GNNs) have demonstrated great success in representation learning for graph-structured data. The layer-wise graph convolution in GNNs is shown to be powerful at capturing graph topology. During this process, GNNs are usually guided by pre-defined kernels such as Laplacian matrix, adjacency matrix, or their variants. However, the adoptions of pre-defined kernels may restrain the generalities to different graphs: mismatch between graph and kernel would entail sub-optimal performance. For example, GNNs that focus on low-frequency information may not achieve satisfactory performance when high-frequency information is significant for the graphs, and vice versa. To solve this problem, in this paper, we propose a novel framework - i.e., namely Adaptive Kernel Graph Neural Network (AKGNN) - which learns to adapt to the optimal graph kernel in a unified manner at the first attempt. In the proposed AKGNN, we first design a data-driven graph kernel learning mechanism, which adaptively modulates the balance between all-pass and low-pass filters by modifying the maximal eigenvalue of the graph Laplacian. Through this process, AKGNN learns the optimal threshold between high and low frequency signals to relieve the generality problem. Later, we further reduce the number of parameters by a parameterization trick and enhance the expressive power by a global readout function. Extensive experiments are conducted on acknowledged benchmark datasets and promising results demonstrate the outstanding performance of our proposed AKGNN by comparison with state-of-the-art GNNs. The source code is publicly available at: https://github.com/jumxglhf/AKGNN.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

Weakly-Supervised Semantic Segmentation (WSSS) segments objects without heavy burden of dense annotation. While as a price, generated pseudo-masks exist obvious noisy pixels, which result in sub-optimal segmentation models trained over these pseudo-masks. But rare studies notice or work on this problem, even these noisy pixels are inevitable after their improvements on pseudo-mask. So we try to improve WSSS in the aspect of noise mitigation. And we observe that many noisy pixels are of high confidences, especially when the response range is too wide or narrow, presenting an uncertain status. Thus, in this paper, we simulate noisy variations of response by scaling the prediction map in multiple times for uncertainty estimation. The uncertainty is then used to weight the segmentation loss to mitigate noisy supervision signals. We call this method URN, abbreviated from Uncertainty estimation via Response scaling for Noise mitigation. Experiments validate the benefits of URN, and our method achieves state-of-the-art results at 71.2% and 41.5% on PASCAL VOC 2012 and MS COCO 2014 respectively, without extra models like saliency detection. Code is available at https://github.com/XMed-Lab/URN.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

We introduce NaturalInversion, a novel model inversion-based method to synthesize images that agrees well with the original data distribution without using real data. In NaturalInversion, we propose: (1) a Feature Transfer Pyramid which uses enhanced image prior of the original data by combining the multi-scale feature maps extracted from the pre-trained classifier, (2) a one-to-one approach generative model where only one batch of images are synthesized by one generator to bring the non-linearity to optimization and to ease the overall optimizing process, (3) learnable Adaptive Channel Scaling parameters which are end-to-end trained to scale the output image channel to utilize the original image prior further. With our NaturalInversion, we synthesize images from classifiers trained on CIFAR-10/100 and show that our images are more consistent with original data distribution than prior works by visualization and additional analysis. Furthermore, our synthesized images outperform prior works on various applications such as knowledge distillation and pruning, demonstrating the effectiveness of our proposed method.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

Understanding the emerging behaviors of reinforcement learning agents may be difficult because such agents are often trained using highly complex and expressive models. In recent years, most approaches developed for explaining agent behaviors rely on domain knowledge or on an analysis of the agent’s learned policy. For some domains, relevant knowledge may not be available or may be insufficient for producing meaningful explanations. We suggest using formal model abstractions and transforms, previously used mainly for expediting the search for optimal policies, to automatically explain discrepancies that may arise between the behavior of an agent and the behavior that is anticipated by an observer. We formally define this problem of Reinforcement Learning Policy Explanation(RLPE), suggest a class of transforms which can be used for explaining emergent behaviors, and suggest meth-ods for searching efficiently for an explanation. We demonstrate the approach on standard benchmarks.

poster session 3 @ blue 4, poster session 7 @ blue 4, poster session 3, poster session 7

Advances in LiDAR sensors provide rich 3D data that supports 3D scene understanding. However, due to occlusion and signal miss, LiDAR point clouds are in practice 2.5D as they cover only partial underlying shapes, which poses a fundamental challenge to 3D perception. To tackle the challenge, we present a novel LiDAR-based 3D object detection model, dubbed Behind the Curtain Detector (BtcDet), which learns the object shape priors and estimates the complete object shapes that are partially occluded (curtained) in point clouds. BtcDet first identifies the regions that are affected by occlusion and signal miss. In these regions, our model predicts the probability of occupancy that indicates if a region contains object shapes and integrates this probability map with detection features and generates high-quality 3D proposals. Finally, the occupancy estimation is integrated into the proposal refinement module to generate accurate bounding boxes. Extensive experiments on the KITTI Dataset and the Waymo Open Dataset demonstrate the effectiveness of BtcDet. Particularly for the 3D detection of both cars and cyclists on the KITTI benchmark, BtcDet surpasses all of the published state-of-the-art methods by remarkable margins. Code is released (https://github.com/Xharlie/BtcDet).

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Current state-of-the-art salieny detection models rely heavily on large datasets of accurate pixel-wise annotations, which cost a lot of time to prepare. There are some weakly supervised methods developed for alleviating the problem, such as image label, bounding box label, and scribble label, while point label still has not to be explored in this field. In this paper, we propose a new point-supervised dataset (P-DUTS) by relabeling the DUTS dataset. In P-DUTS, there is only one labeled point for each salient object. To infer the saliency map, we first design a adaptive masked flood filling algorithm to generate pseudo labels. Then we design a point-supervised saliency detection model based on the transformer to produce the first round of saliency maps. However, we find that due to the sparseness of the label, the weakly supervised model tends to degenerate into a general foreground detection model. To address this issue, we propose a Non-salient suppression (NSS) step to optimize the erroneous sliency maps generated in the first round, and leverage them for the second round of training. Comprehensive experiments on the five largest benchmark datasets demonstrate our method outperforms the previous state-of-the-art methods trained with the stronger supervision and even surpassed several fully supervised state-of-the-art models. The code is available.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

Off-policy sampling and experience replay are key for improving sample efficiency and scaling model-free temporal difference learning methods. When combined with function approximation, such as neural networks, this combination is known as the deadly triad and is potentially unstable. Recently, it has been shown that stability and good performance at scale can be achieved by combining emphatic weightings and multi-step updates. This approach, however, is generally limited to sampling complete trajectories in order, to compute the required emphatic weighting. In this paper we investigate how to combine emphatic weightings with non-sequential, off-line data sampled from a replay buffer. We develop a multi-step emphatic weighting that can be combined with replay, and a time-reversed n-step TD learning algorithm to learn the required emphatic weighting. We show that these state weightings reduce variance compared with prior approaches, while providing convergence guarantees. We tested the approach at scale on Atari 2600 video games, and observed that the newX-ETD(n) agent improved over baseline agents, highlighting both the scalability and broad applicability of our approach.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

This paper introduces a simple yet effective data-centric approach for the task of improving persona-conditioned dialogue agents. Prior model-centric approaches unquestioningly depend on the raw crowdsourced benchmark datasets such as Persona-Chat. In contrast, we aim to fix annotation artifacts in benchmarking, which is orthogonally applicable to any dialogue model. Specifically, we augment relevant personas to improve dialogue dataset/agent, by leveraging the primal-dual structure of the two tasks, predicting dialogue responses and personas based on each other. Experiments on Persona-Chat show that our approach outperforms pre-trained LMs by an 11.7 point gain in terms of accuracy.

poster session 3 @ red 5, poster session 8 @ red 5, oral session 8 @ red 5, poster session 3, poster session 8, oral session 8

Partial Differential Equations (PDEs) are notoriously difficult to solve. In general, closed form solutions are not available and numerical approximation schemes are computationally expensive. In this paper, we propose to approach the solution of PDEs based on a novel technique that combines the advantages of two recently emerging machine learning based approaches. First, physics-informed neural networks (PINNs) learn continuous solutions of PDEs and can be trained with little to no ground truth data. However, PINNs do not generalize well to unseen domains. Second, convolutional neural networks provide fast inference and generalize but either require large amounts of training data or a physics-constrained loss based on finite differences that can lead to inaccuracies and discretization artifacts. We leverage the advantages of both of these approaches by using Hermite spline kernels in order to continuously interpolate a grid-based state representation that can be handled by a CNN. This allows for training without any precomputed training data using a physics-informed loss function only and provides fast, continuous solutions that generalize to unseen domains. We demonstrate the potential of our method at the examples of the incompressible Navier-Stokes equation and the damped wave equation. Our models are able to learn several intriguing phenomena such as Karman vortex streets, the Magnus effect, Doppler effect, interference patterns and wave reflections. Our quantitative assessment and an interactive real-time demo show that we are narrowing the gap in accuracy of unsupervised ML based methods to industrial solvers for computational fluid dynamics (CFD) while being orders of magnitude faster.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

The human intrinsic desire to pursue knowledge, also known as curiosity, is considered essential in the process of skill acquisition. With the aid of artificial curiosity, we could equip current techniques for control, such as Reinforcement Learning, with more natural exploration capabilities. A promising approach in this respect has consisted of using Bayesian surprise on model parameters, i.e. a metric for the difference between prior and posterior beliefs, to favour exploration. In this contribution, we propose to apply Bayesian surprise in a latent space representing the agent’s current understanding of the dynamics of the system, drastically reducing the computational costs. We extensively evaluate our method by measuring the agent's performance in terms of environment exploration, for continuous tasks, and looking at the game scores achieved, for video games. Our model is computationally cheap and compares positively with current state-of-the-art methods on several problems. We also investigate the effects caused by stochasticity in the environment, which is often a failure case for curiosity-driven agents. In this regime, the results suggest that our approach is resilient to stochastic transitions.

poster session 3 @ blue 1, poster session 7 @ blue 1, oral session 7 @ blue 1, poster session 3, poster session 7, oral session 7

Local image feature matching under large appearance, viewpoint, and distance changes is challenging yet important. Conventional methods detect and match tentative local features across the whole images, with heuristic consistency checks to guarantee reliable matches. In this paper, we introduce a novel Overlap Estimation method conditioned on image pairs with TRansformer, named OETR, to constrain local feature matching in the commonly visible region. OETR performs overlap estimation in a two-step process of feature correlation and then overlap regression. As a preprocessing module, OETR can be plugged into any existing local feature detection and matching pipeline, to mitigate potential view angle or scale variance. Intensive experiments show that OETR can boost state-of-the-art local feature matching performance substantially, especially for image pairs with small shared regions. Code will be available upon publication.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Understanding an articulated 3D object with its movable parts is an essential skill for an intelligent agent. This paper presents a novel approach to parse 3D part mobility from point cloud sequences. The key innovation is learning explicit point correspondence from a raw unordered point cloud sequence. We propose a novel deep network called P^3-Net to parallelize trajectory feature extraction and point correspondence establishment, performing joint optimization between them. Specifically, we design a Match-LSTM module to reaggregate point features among different frames by a point correspondence matrix, a.k.a. the matching matrix. To obtain this matrix, an attention module is proposed to calculate the point correspondence. Moreover, we implement a Gumbel-Sinkhorn module to reduce the many-to-one relationship for better point correspondence. We conduct comprehensive evaluations on public benchmarks, including the motion dataset and the PartNet dataset. Results demonstrate that our approach outperforms SOTA methods on various 3D parsing tasks of part mobility, including motion flow prediction, motion part segmentation, and motion attribute (i.e. axis & range) estimation. Moreover, we integrate our approach into a robot perception module to validate its robustness.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Although convolution neural network based stereo matching architectures have made impressive achievements, there are still some limitations: 1) Convolutional Feature (CF) tends to capture appearance information, which is inadequate for accurate matching. 2) Due to the static filters, current convolution based disparity refinement modules often produce over-smooth results. In this paper, we present two schemes to address these issues, where some traditional wisdoms are integrated. Firstly, we introduce a pairwise feature for deep stereo matching networks, named LSP (Local Similarity Pattern). Through explicitly revealing the neighbor relationships, LSP contains rich structural information, which can be leveraged to aid CF for more discriminative feature description. Secondly, we design a dynamic self-reassembling refinement strategy and apply it to the cost distribution and the disparity map respectively. The former could be equipped with the unimodal distribution constraint to alleviate the over-smoothing problem, and the latter is more practical. The effectiveness of the proposed methods is demonstrated via incorporating them into two well-known basic architectures, GwcNet and GANet-deep. Experimental results on the SceneFlow and KITTI benchmarks show that our modules significantly improve the performance of the model.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

While AI planning and Reinforcement Learning (RL) solve sequential decision-making problems, they are based on different formalisms, which leads to a significant difference in their action spaces. When solving planning problems using RL algorithms, we have observed that a naive translation of the planning action space incurs severe degradation in sample complexity. In practice, those action spaces are often engineered manually in a domain-specific manner. In this abstract, we present a method that reduces the parameters of operators in AI planning domains by introducing a parameter seed set problem and casting it as a classical planning task. Our experiment shows that our proposed method significantly reduces the number of actions in the RL environments originating from AI planning domains.

poster session 3 @ blue 4, poster session 7 @ blue 4, poster session 3, poster session 7

Clustering-based unsupervised domain adaptive (UDA) person re-identification (ReID) reduces exhaustive annotations. However, owing to unsatisfactory feature embedding and imperfect clustering, pseudo labels for target domain data inherently contain an unknown proportion of wrong ones, which would mislead feature learning. In this paper, we propose an approach named probabilistic uncertainty guided progressive label refinery (P2LR) for domain adaptive person re-identification. First, we propose to model the labeling uncertainty with the probabilistic distance along with ideal single-peak distributions. A quantitative criterion is established to measure the uncertainty of pseudo labels and facilitate the network training. Second, we explore a progressive strategy for refining pseudo labels. With the uncertainty-guided alternative optimization, we balance between the exploration of target domain data and the negative effects of noisy labeling. On top of a strong baseline, we obtain significant improvements and achieve the state-of-the-art performance on four UDA ReID benchmarks. Specifically, our method outperforms the baseline by 6.5% mAP on the Duke2Market task, while surpassing the state-of-the-art method by 2.5% mAP on the Market2MSMT task.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

We consider a class of visual analogical reasoning problems that involve discovering the sequence of transformations by which pairs of input/output images are related, so as to analogously transform future inputs. This program synthesis task can be easily solved via symbolic search. Using a variation of the ‘neural analogical reasoning’ approach, we instead search for a sequence of elementary neural network transformations that manipulate distributed representations derived from a symbolic space, to which input images are directly encoded. We evaluate the extent to which our ‘neural reasoning’ approach generalises for images with unseen shapes and positions.

poster session 3 @ blue 5, poster session 7 @ blue 5, poster session 3, poster session 7

Spatial and channel attentions, modelling the semantic interdependencies in spatial and channel dimensions respectively, have recently been widely used for semantic segmentation. However, computing spatial and channel attentions separately sometimes causes errors, especially for those difficult cases. In this paper, we propose Channelized Axial Attention (CAA) to seamlessly integrate channel attention and spatial attention into a single operation with negligible computation overhead. Specifically, we break down the dot-product operation of the spatial attention into two parts and insert channel relation in between, allowing for independently optimized channel attention on each spatial location. We further develop grouped vectorization, which allows our model to run with very little memory consumption without slowing down the running speed. Comparative experiments conducted on multiple benchmark datasets, including Cityscapes, PASCAL Context, and COCO-Stuff, demonstrate that our CAA outperforms many state-of-the-art segmentation models (including dual attention) on all tested datasets.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

Knowledge Distillation (KD), which is an effective model compression and acceleration technique, has been successfully applied to graph neural networks (GNNs) recently. Existing approaches utilize a single GNN model as the teacher to distill knowledge. However, we notice that GNN models with different number of layers demonstrate different classification abilities on nodes with different degrees. On the one hand, for nodes with high degrees, their local structures are dense and complex, hence more message passing is needed. Therefore, GNN models with more layers perform better. On the other hand, for nodes with low degrees, whose local structures are relatively sparse and simple, the repeated message passing can easily lead to over-smoothing. Thus, GNN models with less layers are more suitable. However, existing single-teacher GNN knowledge distillation approaches which are based on a single GNN model, are sub-optimal. To this end, we propose a novel approach to distill multi-scale knowledge, which learns from multiple GNN teacher models with different number of layers to capture the topological semantic at different scales. Instead of learning from the teacher models equally, the proposed method automatically assigns proper weights for each teacher model via an attention mechanism which enables the student to select teachers for different local structures. Extensive experiments are conducted to evaluate the proposed method on four public datasets. The experimental results demonstrate the superiority of our proposed method over state-of-the-art methods.

poster session 3 @ blue 3, poster session 8 @ blue 3, oral session 8 @ blue 3, poster session 3, poster session 8, oral session 8

Label-specific features serve as an effective strategy to learn from multi-label data with tailored features accounting for the distinct discriminative properties of each class label. Existing prototype-based label-specific feature transformation approaches work in a three-stage framework, where prototype acquisition, label-specific feature generation and classification model induction are performed independently. Intuitively, this separate framework is suboptimal due to its decoupling nature. In this paper, we make a first attempt towards a unified framework for prototype-based label-specific feature transformation, where the prototypes and the label-specific features are directly optimized for classification. To instantiate it, we propose modelling the prototypes probabilistically by the normalizing flows, which possess adaptive prototypical complexity to fully capture the underlying properties of each class label and allow for scalable stochastic optimization. Then, a label correlation regularized probabilistic latent metric space is constructed via jointly learning the prototypes and the metric-based label-specific features for classification. Comprehensive experiments on 14 benchmark data sets show that our approach outperforms the state-of-the-art counterparts.

poster session 3 @ blue 2, poster session 8 @ blue 2, poster session 3, poster session 8

Unsupervised Salient Object Detection (USOD) is a promising yet challenging task that aims to learn a salient object detection model without any ground-truth labels. Self-supervised learning based methods have achieved remarkable success recently and have become the dominant approach in USOD. However, we observed that two distribution biases of salient objects limit further performance improvement of the USOD methods, namely, contrast distribution bias and spatial distribution bias. Concretely, contrast distribution bias is essentially a confounder that makes images with similar high-level semantic contrast and/or low-level visual appearance contrast spuriously dependent, thus forming data-rich contrast clusters and leading the training process biased towards the data-rich contrast clusters in the data. Spatial distribution bias means that the position distribution of all salient objects in a dataset is concentrated on the center of the image plane, which could be harmful to off-center objects prediction. This paper proposes a causal based debiasing framework to disentangle the model from the impact of such biases. Specifically, we use causal intervention to perform de-confounded model training to minimize the contrast distribution bias and propose an image-level weighting strategy that softly weights each image's importance according to the spatial distribution bias map. Extensive experiments on 6 benchmark datasets show that our method significantly outperforms previous unsupervised state-of-the-art methods and even surpasses some of the supervised methods, demonstrating our debiasing framework's effectiveness.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

Data association is important in the point cloud registration. In this work, we propose to solve the partial-to-partial registration from a new perspective, by introducing multi-level feature interactions between the source and the reference clouds at the feature extraction stage, such that the registration can be realized without the attention or explicit mask estimation for the overlapping detection as adopted previously. Specifically, we present FINet, a feature interaction-based structure with the capability to enable and strengthen the information associating between the inputs at multiple stages. To achieve this, we first split the features into two components, one for rotation and one for translation, based on the fact that they belong to different solution spaces, yielding a dual branches structure. Second, we insert several interaction modules at the feature extractor for the data association. Third, we propose a transformation sensitivity loss to obtain rotation-attentive and translation-attentive features. Experiments demonstrate that our method performs higher precision and robustness compared to the state-of-the-art traditional and learning-based methods. Code is available at https://github.com/megvii-research/FINet.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Group re-identification (G-ReID) focuses on associating the group images containing the same persons under different cameras. The key challenge of G-ReID is that all the cases of the intra-group member and layout variations are hard to exhaust. To this end, we propose a novel uncertainty modeling, which treats each image as a distribution depending on the current member and layout, then digs out potential group features by random samplings. Based on potential and original group features, uncertainty modeling can learn better decision boundaries, which is implemented by two modules, member variation module (MVM) and layout variation module (LVM). Furthermore, we propose a novel second-order transformer framework (SOT), which is inspired by the fact that the position modeling in the transformer is coped with the G-ReID task. SOT is composed of the intra-member module and inter-member module. Specifically, the intra-member module extracts the first-order token for each member, and then the inter-member module learns a second-order token as a group feature by the above first-order tokens, which can be regarded as the token of tokens. A large number of experiments have been conducted on three available datasets, including CSG, DukeGroup and RoadGroup. The experimental results show that the proposed SOT outperforms all previous state-of-the-art methods.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

Question answering over semi-structured tables has attracted significant attention in the NLP community. However, most of the existing work focus on questions that can be answered with short-form answer, i.e. the answer is often a table cell or aggregation of multiple cells. This can mismatch with the intents of users who want to ask more complex questions that require free-form answers such as explanations. To bridge the gap, most recently, pre-trained sequence-to-sequence language models such as T5 are used for generating free-form answers based on the question and table inputs. However, these pre-trained language models have weaker encoding abilities over table cells and schema. To mitigate this issue, in this work, we present an intermediate pre-training framework, Generation-focused Table-based Intermediate Pre-training (GENTAP), that jointly learns representations of natural language questions and tables. GENTAP learns to generate via two training objectives to enhance the question understanding and table representation abilities for complex questions. Based on experimental results, models that leverage GENTAP framework outperform the existing baselines on FETAQA benchmark. The pre-trained models are not only useful for free-form question answering, but also for few-shot data-to-text generation task, thus showing good transfer ability by obtaining new state-of-the-art results.

poster session 3 @ red 4, poster session 12 @ red 4, oral session 12 @ red 4, poster session 3, poster session 12, oral session 12

Homogeneous instance segmentation aims to identify each instance in an image where all interested instances belong to the same category, such as plant leaves and microscopic cells. Recently, proposal-free methods have received increasing attention due to their efficient pipeline, which straightforwardly generate instance-aware information to group pixels into different instances. However, they often fail to distinguish adjacent instances due to similar appearance, dense distribution and ambiguous boundaries of instances in homogeneous images. In this paper, we propose a pixel-embedded affinity modeling method for homogeneous instance segmentation, which is able to preserve the semantic information of instances and improve the distinguishability of adjacent instances. Instead of predicting affinity directly, we propose a self-correlation module to explicitly model the pairwise relationships between pixels, by estimating the similarity between embeddings generated from the input image through CNNs. Based on the self-correlation module, we further design a cross-correlation module to maintain the semantic consistency between instances. Specifically, we map the transformed input images with different views and appearances into the same embedding space, and then mutually estimate the pairwise relationships of embeddings generated from the original input and its transformed variants. In addition, to integrate the global instance information, we introduce an embedding pyramid module to model affinity on different scales. Extensive experiments demonstrate the versatile and superior performance of our method on three representative datasets.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

Named Entity Recognition models perform well on benchmark datasets but fail to generalize well even in the same domain. The goal of my th esis is to quantify the degree of in-domain generalization in NER, probe models for entity name vs. context learning and finally improve their robustness, focusing on the recognition of ethnically diverse entities and new entities over time when the models are deployed.

poster session 3 @ red 5, poster session 1 @ red 5, poster session 3, poster session 1

Understanding how a city’s physical appearance and environmental surroundings impact society traits, such as safety, is an essential issue in social artificial intelligence. To demonstrate the relationship, most existing studies utilize subjective human perceptual attributes, categorization only for a few violent crimes, and images taken from still shot images. These lead to difficulty in identifying location-specific characteristics for urban safety. In this work, to address this problem, we propose a large-scale dataset and a novel method by adopting a concept of “Deviance" which explains behaviors violating social norms, both formally (e.g. crime) and informally (e.g. civil complaints). We first collect a geo-tagged dataset consisting of incident report data for seven metropolitan cities, with corresponding sequential images around incident sites obtained from Google street view. We also design a convolutional neural network that learns spatio-temporal visual attributes of deviant streets. Experimental results show that our framework can reliably recognize real-world deviance in various cities. Furthermore, we analyze which visual attribute is important for deviance identification and severity estimation. We have released our dataset and source codes at our project page: https://deviance-project.github.io/DevianceNet/

poster session 3 @ red 6, poster session 12 @ red 6, poster session 3, poster session 12

Multi-agent reinforcement learning (MARL) provides an efficient way for simultaneously learning policies for multiple agents interacting with each other. However, in scenarios requiring complex interactions, existing algorithms can suffer from an inability to accurately anticipate the influence of self-actions on other agents. Incorporating an ability to reason about other agents' potential responses can allow an agent to formulate more effective strategies. This paper adopts a recursive reasoning model in a centralized-training-decentralized-execution framework to help learning agents better cooperate with or compete against others. The proposed algorithm, referred to as the Recursive Reasoning Graph (R2G), shows state-of-the-art performance on multiple multi-agent particle and robotics games.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

Traditional symbolic reasoning engines, while attractive for their precision and explicability, have a few major drawbacks: the use of brittle inference procedures that rely on exact matching (unification) of logical terms, an inability to deal with uncertainty, and the need for a precompiled rule-base of knowledge (the “knowledge acquisition” problem). To address these issues, we devise a novel logical reasoner called Braid, that supports probabilistic rules, and uses the notion of custom unification functions and dynamic rule generation to overcome the brittle matching and knowledge-gap problem prevalent in traditional reasoners. In this paper, we describe the reasoning algorithms used in Braid, and their implementation in a distributed task-based framework that builds proof/explanation graphs for an input query. We use a simple QA example from a children’s story to motivate Braid’s design and explain how the various components work together to produce a coherent logical explanation. Finally, we evaluate Braid on the ROC Story Cloze test and achieve close to state-of-the-art results while providing frame-based explanations.

poster session 3 @ blue 5, poster session 7 @ blue 5, poster session 3, poster session 7

Recently, explanation methods have been proposed to evaluate the predictions of Graph Neural Networks on the task of link prediction. Evaluating explanation quality is difficult without ground truth explanations. This thesis is focused on providing a method, including datasets and scoring metrics, to quantitatively evaluate explanation methods on link prediction on Knowledge Graphs.

poster session 3 @ blue 5, poster session 1 @ blue 5, poster session 3, poster session 1

Modern Convolutional Neural Network (CNN) architectures, despite their superiority in solving various problems, are generally too large to be deployed on resource constrained edge devices. In this paper, we reduce memory usage and floating-point operations required by convolutional layers in CNNs. We compress these layers by generalizing the Kronecker Product Decomposition to apply to multidimensional tensors, leading to the Generalized Kronecker Product Decomposition} (GKPD). Our approach yields a plug-and-play module that can be used as a drop-in replacement for any convolutional layer. Experimental results for image classification on CIFAR-10 and ImageNet datasets using ResNet, MobileNetv2 and SeNet architectures substantiate the effectiveness of our proposed approach. We find that GKPD outperforms state-of-the-art decomposition methods including Tensor-Train and Tensor-Ring as well as other relevant compression methods such as pruning and knowledge distillation.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

For users to trust planning algorithms, they must be able to understand the planner's outputs and the reasons for each action selection. This output does not tend to be user-friendly, often consisting of sequences of parametrised actions or task networks. And these may not be practical for non-expert users who may find it easier to read natural language descriptions. In this paper, we propose PlanVerb, a domain and planner-independent method for the verbalization of task plans. It is based on semantic tagging of actions and predicates. Our method can generate natural language descriptions of plans including causal explanations. The verbalized plans can be summarized by compressing the actions that act on the same parameters. We further extend the concept of verbalization space, previously applied to robot navigation, and apply it to planning to generate different kinds of plan descriptions for different user requirements. Our method can deal with PDDL and RDDL domains, provided that they are tagged accordingly. Our user survey evaluation shows that users can read our automatically generated plan descriptions and that the explanations help them answer questions about the plan.

poster session 3 @ blue 4, poster session 12 @ blue 4, poster session 3, poster session 12

To deal with ambiguities in partial multi-label learning (PML), existing popular PML research attempts to perform disambiguation by direct ground-truth label identification. However, these approaches can be easily misled by noisy false-positive labels in the iteration of updating the model parameter and the latent ground-truth label variables. When labeling information is ambiguous, we should depend more on underlying structure of data, such as label and feature correlations, to perform disambiguation for partially labeled data. Moreover, large margin nearest neighbour (LMNN) is a popular strategy that considers data structure in classification. However, due to the ambiguity of labeling information in PML, traditional LMNN cannot be used to solve the PML problem directly. In addition, embedding is an effective technology to decrease the noise information of data. Inspried by LMNN and embedding technology, we propose a novel PML paradigm called Partial Multi-label Learning via Large Margin Nearest Neighbour Embeddings (PML-LMNNE), which aims to conduct disambiguation by projecting labels and features into a lower-dimension embedding space and reorganize the underlying structure by LMNN in the embedding space simultaneously. An efficient algorithm is designed to implement the proposed method and the convergence rate of the algorithm is analyzed. Moreover, we present a theoretical analysis of the generalization error bound for the proposed PML-LMNNE, which shows that the generalization error converges to the sum of two times the Bayes error over the labels when the number of instances $n$ goes to infinity. Comprehensive experiments on artificial and real-world datasets demonstrate the superiorities of the proposed PML-LMNNE.

poster session 3 @ blue 2, poster session 8 @ blue 2, oral session 3 @ blue 2, poster session 3, poster session 8, oral session 3

Recently, graph neural networks (GNNs) have been widely used for document classification. However, most existing methods are based on static word co-occurrence graphs without sentence-level information, which poses three challenges:(1) word ambiguity, (2) word synonymity, and (3) dynamic contextual dependency. To address these challenges, we propose a novel GNN-based sparse structure learning model for inductive document classification. Specifically, a document-level graph is initially generated by a disjoint union of sentence-level word co-occurrence graphs. Our model collects a set of trainable edges connecting disjoint words between sentences, and employs structure learning to sparsely select edges with dynamic contextual dependencies. Graphs with sparse structure can jointly exploit local and global contextual information in documents through GNNs.For inductive learning, the refined document graph is further fed into a general readout function for graph-level classification and optimization in an end-to-end manner. Extensive experiments on several real-world datasets demonstrate that the proposed model outperforms most state-of-the-art results, and reveal the necessity to learn sparse structures for each document.

poster session 3 @ red 4, poster session 12 @ red 4, poster session 3, poster session 12

Simultaneous Localization and Mapping (SLAM) and Autonomous Driving are becoming increasingly more important in recent years. Point cloud-based large scale place recognition is the spine of them. While many models have been proposed and have achieved acceptable performance by learning short-range local features, they always skip long-range contextual properties. Moreover, the model size also becomes a serious shackle for their wide applications. To overcome these challenges, we propose a super light-weight network model termed SVT-Net. On top of the highly efficient 3D Sparse Convolution (SP-Conv), an Atom-based Sparse Voxel Transformer (ASVT) and a Cluster-based Sparse Voxel Transformer (CSVT) are proposed respectively to learn both short-range local features and long-range contextual features. Consisting of ASVT and CSVT, SVT-Net can achieve state-of-the-art performance in terms of both recognition accuracy and running speed with a super-light model size (0.9M parameters). Meanwhile, for the purpose of further boosting efficiency, we introduce two simplified versions, which also achieve state-of-the-art performance and further reduce the model size to 0.8M and 0.4M respectively.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Contextual word representations generated by language models learn spurious associations present in the training corpora. Adversaries can exploit these associations to reverse-engineer the private attributes of entities mentioned in the training corpora. These findings have led to efforts towards minimizing the privacy risks of language models. However, existing approaches lack interpretability, compromise on data utility and fail to provide privacy guarantees. Thus, the goal of my doctoral research is to develop interpretable approaches towards privacy preservation of text representations that maximize data utility retention and guarantee privacy. To this end, I aim to study and develop methods to incorporate steganographic modifications within the vector geometry to obfuscate underlying spurious associations and retain the distributional semantic properties learnt during training.

poster session 3 @ red 5, poster session 1 @ red 5, poster session 3, poster session 1

Recent studies have shown that strong Natural Language Understanding (NLU) models are prone to relying on annotation biases of the datasets as a shortcut, which goes against the underlying mechanisms of the task of interest. To reduce such biases, several recent works introduce debiasing methods to regularize the training process of targeted NLU models. In this paper, we provide a new perspective with causal inference to find out the bias. On one hand, we show that there is an unobserved confounder for the natural language utterances and their respective classes, leading to spurious correlations from training data. To remove such confounder, the backdoor adjustment with causal intervention is utilized to find the true causal effect, which makes the training process fundamentally different from the traditional likelihood estimation. On the other hand, in inference process, we formulate the bias as the direct causal effect and remove it by pursuing the indirect causal effect with counterfactual reasoning. We conduct experiments on large-scale natural language inference and fact verification benchmarks, evaluating on bias sensitive datasets that are specifically designed to assess the robustness of models against known biases in the training data. Experimental results show that our proposed debiasing framework outperforms previous state-of-the-art debiasing methods while maintaining the original in-distribution performance.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

In the Bayesian reinforcement learning (RL) setting, a prior distribution over the unknown problem parameters -- the rewards and transitions -- is assumed, and a policy that optimizes the (posterior) expected return is sought. A common approximation, which has been recently popularized as meta-RL, is to train the agent on a sample of N problem instances from the prior, with the hope that for large enough N, good generalization behavior to an unseen test instance will be obtained. In this work, we study generalization in Bayesian RL under the probably approximately correct (PAC) framework, using the method of algorithmic stability. Our main contribution is showing that by adding regularization, the optimal policy becomes uniformly stable in an appropriate sense. Most stability results in the literature build on strong convexity of the regularized loss -- an approach that is not suitable for RL as Markov decision processes (MDPs) are not convex. Instead, building on recent results of fast convergence rates for mirror descent in regularized MDPs, we show that regularized MDPs satisfy a certain quadratic growth criterion, which is sufficient to establish stability. This result, which may be of independent interest, allows us to study the effect of regularization on generalization in the Bayesian RL setting.

poster session 3 @ blue 1, poster session 7 @ blue 1, oral session 7 @ blue 1, poster session 3, poster session 7, oral session 7

We introduce a novel technique to identify three spectra representing the three primary materials in a hyperspectral image of a scene. We accomplish this using a modified autoencoder. Further research will be conducted to verify the accuracy of these spectra.

poster session 3 @ red 4, poster session 4 @ red 4, poster session 3, poster session 4

The high efficiency in computation and storage makes hashing (including binary hashing and quantization) a common strategy in large-scale retrieval systems. To alleviate the reliance on expensive annotations, unsupervised deep hashing becomes an important research problem. This paper provides a novel solution to unsupervised deep quantization, namely Contrastive Quantization with Code Memory (MeCoQ). Different from existing reconstruction-based strategies, we learn unsupervised binary descriptors by contrastive learning, which can better capture discriminative visual semantics. Besides, we uncover that codeword diversity regularization is critical to prevent contrastive learning-based quantization from model degeneration. Moreover, we introduce a novel quantization code memory module that boosts contrastive learning with lower feature drift than conventional feature memories. Extensive experiments on benchmark datasets show that MeCoQ outperforms state-of-the-art methods.

poster session 3 @ red 1, poster session 8 @ red 1, oral session 3 @ red 1, poster session 3, poster session 8, oral session 3

This paper proposes a new sequential model learning architecture to solve partially observable Markov decision problems. Rather than compressing sequential information at every timestep as in conventional recurrent neural network-based methods, the proposed architecture generates a latent variable in each data block with a length of multiple timesteps and passes the most relevant information to the next block for policy optimization. The proposed blockwise sequential model is implemented based on self-attention, making the model capable of detailed sequential learning in partial observable settings. The proposed model builds an additional learning network to efficiently implement gradient estimation by using self-normalized importance sampling, which does not require the complex blockwise input data reconstruction in the model learning. Numerical results show that the proposed method significantly outperforms previous methods in various partially observable environments.

poster session 3 @ blue 1, poster session 7 @ blue 1, oral session 7 @ blue 1, poster session 3, poster session 7, oral session 7

Unsupervised/self-supervised pre-training methods for graph representation learning have recently attracted increasing research interests, and they are shown to be able to generalize to various downstream applications. Yet, the adversarial robustness of such pre-trained graph learning models remains largely unexplored. More importantly, most existing defense techniques designed for end-to-end graph representation learning methods require pre-specified label definitions, and thus cannot be directly applied to the pre-training methods. In this paper, we propose an unsupervised defense technique to robustify pre-trained deep graph models, so that the perturbations on the input graph can be successfully identified and blocked before the model is applied to different downstream tasks. Specifically, we introduce a mutual information-based measure, graph representation vulnerability (GRV), to quantify the robustness of graph encoders on the representation space. We then formulate an optimization problem to learn the graph representation by carefully balancing the trade-off between the expressive power and the robustness (i.e., GRV) of the graph encoder. The discrete nature of graph topology and the joint space of graph data make the optimization problem intractable to solve. To handle the above difficulty and to reduce computational expense, we further relax the problem and thus provide an approximate solution. Additionally, we explore a provable connection between the robustness of the unsupervised graph encoder and that of models on downstream tasks. Extensive experiments demonstrate that even without access to labels and tasks, our model is still able to enhance robustness against adversarial attacks on three downstream tasks (node classification, link prediction, and community detection) by an average of +16.5\% compared with existing methods. Our codes are available at https://github.com/anonymous-Robustgraph/Robustgraph.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

Person search has recently emerged as a challenging task that jointly addresses pedestrian detection and person re-identification. Existing approaches follow a fully supervised setting where both bounding box and identity annotations are available. However, annotating identities is labor-intensive, limiting the practicability and scalability of current frameworks. This paper inventively considers weakly supervised person search with only bounding box annotations. We propose to address this novel task by investigating three levels of context clues (i.e., detection, memory and scene) in unconstrained natural images. The first two are employed to promote local and global discriminative capabilities, while the latter enhances clustering accuracy. Despite its simple design, our CGPS boosts the baseline model by 8.8% in mAP on CUHK-SYSU. Surprisingly, it even achieves comparable performance with several supervised person search models. The codes are available in the supplementary materials and will be released.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

We present a simple and concise semantics for temporal planning. Our semantics are developed and formalised in the logic of the interactive theorem prover Isabelle/HOL. We derive from those semantics a validation algorithm for temporal planning and show, using a formal proof in Isabelle/HOL, that this validation algorithm implements our semantics. We experimentally evaluate our verified validation algorithm and show that it performs reasonably well.

poster session 3 @ blue 4, poster session 12 @ blue 4, poster session 3, poster session 12

Unsupervised domain adaptive person re-identification aims at learning on an unlabeled target domain with only labeled data in source domain. Currently, the state-of-the-arts usually solve this problem by pseudo-label-based clustering and fine-tuning in target domain. However, the reason behind the noises of pseudo labels is not sufficiently explored, especially for the popular multi-branch models. We argue that the consistency between different feature spaces is the key to the pseudo labels’ quality. Then a \textbf{SE}lf-\textbf{C}onsistent pseudo label \textbf{R}efin\textbf{E}men\textbf{T} method, termed as SECRET, is proposed to improve consistency by mutually refining the pseudo labels generated from different feature spaces. The proposed SECRET gradually encourages the improvement of pseudo labels’ quality during training process, which further leads to better cross-domain Re-ID performance. Extensive experiments on benchmark datasets show the superiority of our method. Specifically, our method outperforms the state-of-the-arts by $\mathbf{6.3\%}$ in terms of mAP on the challenging dataset MSMT17. In the purely unsupervised setting, our method also surpasses existing works by a large margin. Code is available at \url{https://github.com/LunarShen/SECRET}.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

A key trait of daily conversations between individuals is the ability to express empathy towards others, and exploring ways to implement empathy is a crucial step towards human-like dialogue systems. Previous approaches on this topic mainly focus on detecting and utilizing the user's emotion for generating empathetic responses. However, since empathy includes both aspects of affection and cognition, we argue that in addition to identifying the user's emotion, cognitive understanding of the user's situation should also be considered. To this end, we propose a novel approach for empathetic response generation, which leverages commonsense to draw more information about the user's situation and uses this additional information to further enhance the empathy expression in generated responses. We evaluate our approach on EmpatheticDialogues, which is a widely-used benchmark dataset for empathetic response generation. Empirical results demonstrate that our approach outperforms the baseline models in both automatic and human evaluations and can generate more informative and empathetic responses.

poster session 3 @ red 5, poster session 8 @ red 5, oral session 3 @ red 5, poster session 3, poster session 8, oral session 3

Unsupervised point cloud registration algorithm usually suffers from the unsatisfied registration precision in the partially overlapping problem due to the lack of effective inlier evaluation. In this paper, we propose a neighborhood consensus based reliable inlier evaluation method for robust unsupervised point cloud registration. It is expected to capture the discriminative geometric difference between the source neighborhood and the corresponding pseudo target neighborhood for effective inlier distinction. Specifically, our model consists of a matching map refinement module and an inlier evaluation module. In our matching map refinement module, we improve the point-wise matching map estimation by integrating the matching scores of neighbors into it. The aggregated neighborhood information potentially facilitates the discriminative map construction so that high-quality correspondences can be provided for generating the pseudo target point cloud. Based on the observation that the outlier has the significant structure-wise difference between its source neighborhood and corresponding pseudo target neighborhood while this difference for inlier is small, the inlier evaluation module exploits this difference to score the inlier confidence for each estimated correspondence. In particular, we construct an effective graph representation for capturing this geometric difference between the neighborhoods. Finally, with the learned correspondences and the corresponding inlier confidence, we use the weighted SVD algorithm for transformation estimation.Under the unsupervised setting, we exploit the Huber function based global alignment loss, the local neighborhood consensus loss and spatial consistency loss for model optimization. The experimental results on extensive datasets demonstrate that our unsupervised point cloud registration method can yield comparable performance.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

An argument can be regarded as some premises and a claim following from those premises. Normally, arguments exchanged by human agents are enthymemes, which generally means that some premises are implicit. So when an enthymeme is presented, the presenter expects that the recipient can identify the missing premises. An important kind of implicitness arises when a presenter assumes that two symbols denote the same, or nearly the same, concept (e.g. dad and father), and uses the symbols interchangeably. To model this process, we propose the use of semantic distance measures (e.g. based on a vector representation of word embeddings or a semantic network representation of words) to determine whether one symbol can be substituted by another. We present a theoretical framework for using substitutions, together with abduction of default knowledge, for understanding enthymemes based on deductive argumentation, and investigate how this could be used in practice.

poster session 3 @ blue 5, poster session 7 @ blue 5, poster session 3, poster session 7

The deployment of systems of artificial intelligence (AI) in high-risk settings warrants the need for trustworthy AI. This crucial requirement is highlighted by recent EU guidelines and regulations, but also by recommendations from OECD and UNESCO, among several other examples. One critical premise of trustworthy AI involves the necessity of finding explanations that offer reliable guarantees of soundness. This paper argues that the best known eXplainable AI (XAI) approaches fail to provide sound explanations, or that alternatively find explanations which can exhibit significant redundancy. The solution to these drawbacks are explanation approaches that offer formal guarantees of rigor. These formal explanations are not only sound but guarantee irredundancy. This paper summarizes the recent developments in the emerging discipline of formal XAI. The paper also outlines existing challenges for formal XAI.

poster session 3 @ blue 6, poster session 7 @ blue 6, oral session 3 @ blue 6, poster session 3, poster session 7, oral session 3

Clustering is important for domain adaptive person re-identification(re-ID). A majority of unsupervised domain adaptation (UDA) methods conduct clustering on the target domain and then use the generated pseudo labels for adaptive training. Albeit important, the clustering pipeline adopted by current literature is quite standard and lacks consideration for two characteristics of re-ID, i.e., 1) a single person has various feature distribution in multiple cameras. 2) a person’s occurrence in the same camera are usually temporally continuous. We argue that the multi-camera distribution hinders clustering because it enlarges the intra-class distances. In contrast, the temporal continuity prior is beneficial, because it offers clue for distinguishing some look-alike person (who are temporally far away from each other). These two insight motivate us to propose a novel Divide-And-Regroup Clustering (DARC) pipeline for re-ID UDA. Specifically, DARC divides the unlabeled data into multiple camera-specific groups and conducts local clustering within each camera. Afterwards, it regroups those local clusters potentially belonging to the same person into a unity. Through this divide-and-regroup pipeline, DARC avoids directly clustering across multiple cameras and focuses on the feature distribution within each individual camera. Moreover, during the local clustering, DARC uses the temporal continuity prior to distinguish some look-alike person and thus reduces false positive pseudo labels. Consequentially, DARC effectively reduces clustering errors and improves UDA. Importantly, we show that DARC is compatible to many pseudo label-based UDA methods and brings general improvement. Based on a recent UDA method, DARC advances the state of the art (e.g, 85.1% mAP on MSMT-to-Market and 83.1% mAP on PersonX-to-Market).

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

Federated learning (FL) enables multiple clients to collaboratively train a globally generalized model while keeping local data decentralized. A key challenge in FL is to handle the heterogeneity of data distributions among clients. The local model will shift the global feature when fitting local data, which results in forgetting the global knowledge. Following the idea of knowledge distillation, the global model's prediction can be utilized to help local models preserve the global knowledge in FL. However, when the global model hasn't converged completely, its predictions tend to be less reliable on certain classes, which may results in distillation's misleading of local models. In this paper, we propose a class-wise adaptive self distillation (FedCAD) mechanism to ameliorate this problem. We design class-wise adaptive terms to soften the influence of distillation loss according to the global model's performance on each class and therefore avoid the misleading. Experiments show that our method outperforms other state-of-the-art FL algorithms on benchmark datasets.

poster session 3 @ blue 5, poster session 7 @ blue 5, poster session 3, poster session 7

This paper focuses on a dynamic aspect of responsible autonomy, namely, to make intelligent agents be responsible at run time. That is, it considers settings where decision making by agents impinges upon the outcomes perceived by other agents. For an agent to act responsibly, it must accommodate the desires and other attitudes of its users and, through other agents, of their users. The contribution of this paper is twofold. First, it provides a conceptual analysis of consent, its benefits and misuses, and how understanding consent can help achieve responsible autonomy. Second, it outlines challenges for AI (in particular, for agents and multiagent systems) that merit investigation to form as a basis for modeling consent in multiagent systems and applying consent to achieve responsible autonomy.

poster session 3 @ blue 6, poster session 7 @ blue 6, oral session 7 @ blue 6, poster session 3, poster session 7, oral session 7

Cooperative multi-agent reinforcement learning (MARL) faces significant scalability issues due to state and action spaces that are exponentially large in the number of agents. As environments grow in size, effective credit assignment becomes increasingly harder and often results in infeasible learning times. Still, in many real-world settings, there exist simplified underlying dynamics that can be leveraged for more scalable solutions. In this work, we exploit such locality structures effectively whilst maintaining global cooperation. We propose a novel, value-based multi-agent algorithm called LOMA$Q$, which incorporates local rewards in the Centralized Training Decentralized Execution paradigm. Additionally, we provide a direct reward decomposition method for finding these local rewards when only a global signal is provided. We test our method empirically, showing it scales well compared to other methods, significantly improving performance and convergence speed.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

In order to fully perceive the surrounding environment, many intelligent robots and self-driving cars are equipped with a multi-camera system. Based on this system, the structure-from-motion (SfM) technology is used to realize scene reconstruction, but the fixed relative poses between cameras in the multi-camera system are usually not considered. This paper presents a tailor-made multi-camera based motion averaging system, where the fixed relative poses are utilized to improve the accuracy and robustness of SfM. Our approach starts by dividing the images into reference images and non-reference images, and edges in view-graph are divided into four categories accordingly. Then, a multi-camera based rotating averaging problem is formulated and solved in two stages, where an iterative re-weighted least squares scheme is used to deal with outliers. Finally, a multi-camera based translation averaging problem is formulated and a l1-norm based optimization scheme is proposed to compute the relative translations of multi-camera system and reference camera positions simultaneously. Experiments demonstrate that our algorithm achieves superior accuracy and robustness on various data sets compared to the state-of-the-art methods.

poster session 3 @ red 2, poster session 8 @ red 2, oral session 8 @ red 2, poster session 3, poster session 8, oral session 8

The goal of building intelligent dialogue systems has largely been separately pursued under two paradigms: task-oriented dialogue (TOD) systems, which perform task-specific functions, and open-domain dialogue (ODD) systems, which focus on non-goal-oriented chitchat. The two dialogue modes can potentially be intertwined together seamlessly in the same conversation, as easily done by a friendly human assistant. Such ability is desirable in conversational agents, as the integration makes them more accessible and useful. Our paper addresses this problem of fusing TODs and ODDs in multi-turn dialogues. Based on the popular TOD dataset MultiWOZ, we build a new dataset FusedChat, by rewriting the existing TOD turns and adding new ODD turns. This procedure constructs conversation sessions containing exchanges from both dialogue modes. It features inter-mode contextual dependency, i.e., the dialogue turns from the two modes depend on each other. Rich dependency patterns such as co-reference and ellipsis are included. The new dataset, with 60k new human-written ODD turns and 5k re-written TOD turns, offers a benchmark to test a dialogue model's ability to perform inter-mode conversations. This is a more challenging task since the model has to determine the appropriate dialogue mode and generate the response based on the inter-mode context. However, such models would better mimic human-level conversation capabilities. We evaluate two baseline models on this task, including the classification-based two-stage models and the two-in-one fused models. We publicly release FusedChat (https://github.com/tomyoung903/FusedChat) and the baselines to propel future work on inter-mode dialogue systems.

poster session 3 @ red 5, poster session 8 @ red 5, oral session 8 @ red 5, poster session 3, poster session 8, oral session 8

A conversation is a sequence of utterances, where each utterance plays a specific discourse role while expressing a particular emotion. This paper proposes a novel method to exploit latent discourse role information of an utterance to determine the emotion it conveys in a conversation. Specifically, we use a variant of the Variational-Autoencoder (VAE) to model the context-aware latent discourse roles of each utterance in an unsupervised way. The latent discourse role representation further equips the utterance representation with a salient clue for more accurate emotion recognition. Our experiments show that our proposed method beats the best-reported performances on three public Emotion Recognition in Conversation datasets. This proves that the discourse role information of an utterance plays an important role in the emotion recognition task, which no previous work has studied.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Federated learning (FL) enables distribution of machine learning workloads from the cloud to resource-limited edge devices. Unfortunately, current deep networks remain not only too compute-heavy for inference and training on edge devices, but also too large for communicating updates over bandwidth-constrained networks. In this paper, we develop, implement, and experimentally validate a novel FL framework termed Federated Dynamic Sparse Training (FedDST) by which complex neural networks can be deployed and trained with substantially improved efficiency in both on-device computation and in-network communication. At the core of FedDST is a dynamic process that extracts and trains sparse sub-networks from the target full network. With this scheme, "two birds are killed with one stone:'' instead of full models, each client performs efficient training of its own sparse networks, and only sparse networks are transmitted between devices and the cloud. Furthermore, our results reveal that the dynamic sparsity during FL training more flexibly accommodates local heterogeneity in FL agents than the fixed, shared sparse masks. Moreover, dynamic sparsity naturally introduces an "in-time self-ensembling effect'' into the training dynamics, and improves the FL performance even over dense training. In a realistic and challenging non i.i.d. FL setting, FedDST consistently outperforms competing algorithms in our experiments: for instance, at any fixed upload data cap on non-iid CIFAR-10, it gains an impressive accuracy advantage of 10% over FedAvgM when given the same upload data cap; the accuracy gap remains 3% even when FedAvgM is given 2 times the upload data cap, further demonstrating efficacy of FedDST. Code is available at: https://github.com/bibikar/feddst.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

Few-shot text classification has recently been promoted by the meta-learning paradigm which aims to identify target classes with knowledge transferred from source classes with sets of small tasks named episodes. Despite their success, existing works building their meta-learner based on Prototypical Networks are unsatisfactory in learning discriminative text representations between similar classes, which may lead to contradictions during label prediction. In addition, the task-level and instance-level overfitting problems in few-shot text classification caused by a few training examples are not sufficiently tackled. In this work, we propose a contrastive learning framework named ContrastNet to tackle both discriminative representation and overfitting problems in few-shot text classification. ContrastNet learns to pull closer text representations belonging to the same class and push away text representations belonging to different classes, while simultaneously introducing unsupervised contrastive regularization at both task-level and instance-level to prevent overfitting. Experiments on 8 few-shot text classification datasets show that ContrastNet outperforms the current state-of-the-art models.

poster session 3 @ red 4, poster session 12 @ red 4, poster session 3, poster session 12

Heterogeneity across clients in federated learning (FL) usually hinders the optimization convergence and generalization performance when the aggregation of clients' knowledge occurs in the gradient space. For example, clients may differ in terms of data distribution, network latency, input/output space, and/or model architecture, which can easily lead to the misalignment of their local gradients. To improve the tolerance to heterogeneity, we propose a novel federated prototype learning (FedProto) framework in which the clients and server communicate the abstract class prototypes instead of the gradients. FedProto aggregates the local prototypes collected from different clients, and then sends the global prototypes back to all clients to regularize the training of local models. The training on each client aims to minimize the classification error on the local data while keeping the resulting local prototypes sufficiently close to the corresponding global ones. Moreover, we provide a theoretical analysis to the convergence rate of FedProto under non-convex objectives. In experiments, we propose a benchmark setting tailored for heterogeneous FL, with FedProto outperforming several recent FL approaches on multiple datasets.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

In this paper, we present the texture reformer, a fast and universal neural-based framework for interactive texture transfer with user-specified guidance. The challenges lie in three aspects: 1) the diversity of tasks, 2) the simplicity of guidance maps, and 3) the execution efficiency. To address these challenges, our key idea is to use a novel feed-forward multi-view and multi-stage synthesis procedure consisting of I) a global view structure alignment stage, II) a local view texture refinement stage, and III) a holistic effect enhancement stage to synthesize high-quality results with coherent structures and fine texture details in a coarse-to-fine fashion. In addition, we also introduce a novel learning-free view-specific texture reformation (VSTR) operation with a new semantic map guidance strategy to achieve more accurate semantic-guided and structure-preserved texture transfer. The experimental results on a variety of application scenarios demonstrate the effectiveness and superiority of our framework. And compared with the state-of-the-art interactive texture transfer algorithms, it not only achieves higher quality results but, more remarkably, also is 2-5 orders of magnitude faster. Code is available at https://github.com/EndyWon/Texture-Reformer.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

Unsupervised domain adaptation (UDA) has been highly successful in transferring knowledge acquired from a label-rich source domain to a label-scarce target domain. Open-set domain adaptation (open-set DA) and universal domain adaptation (UniDA) have been proposed as solutions to the problem concerning the presence of additional novel categories in the target domain. Existing open-set DA and UniDA approaches treat all novel categories as one unified unknown class and attempt to detect this unknown class during the training process. However, the features of the novel categories learned by these methods are not discriminative. This limits the applicability of UDA in the further classification of these novel categories into their original categories, rather than assigning them to a single unified class. In this paper, we propose a self-labeling framework to cluster all target samples, including those in the ``unknown'' categories. We train the network to learn the representations of target samples via self-supervised learning (SSL) and to identify the seen and unseen (novel) target-sample categories simultaneously by maximizing the mutual information between labels and input data. We evaluated our approach under different DA settings and concluded that our method generally outperformed existing ones by a wide margin.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

Despite the super-human accuracy of recent deep models in NLP tasks, their robustness is reportedly limited due to their reliance on spurious patterns. We thus aim to leverage contrastive learning and counterfactual augmentation for robustness. For augmentation, existing work either requires humans to add counterfactuals to the dataset or machines to automatically matches near-counterfactuals already in the dataset. Unlike existing augmentation is affected by spurious correlations, ours, by synthesizing “a set” of counterfactuals, and making a collective decision on the distribution of predictions on this set, can robustly supervise the causality of each term. Our empirical results show that our approach, by collective decisions, is less sensitive to task model bias of attribution-based synthesis, and thus achieves significant improvements, in diverse dimensions: 1) counterfactual robustness, 2) cross-domain generalization, and 3) generalization from scarce data.

poster session 3 @ red 4, poster session 12 @ red 4, poster session 3, poster session 12

Symmetry and dominance breaking can be crucial for solving hard combinatorial search and optimisation problems, but the correctness of these techniques sometimes relies on subtle arguments. For this reason, it is desirable to produce efficient, machine-verifiable certificates that solutions have been computed correctly. Building on the cutting planes proof system, we develop a certification method for optimisation problems in which symmetry and dominance breaking are easily expressible. Our experimental evaluation demonstrates that we can efficiently verify fully general symmetry breaking in Boolean satisfiability (SAT) solving, thus providing, for the first time, a unified method to certify a range of advanced SAT techniques that also includes XOR and cardinality reasoning. In addition, we apply our method to maximum clique solving and constraint programming as a proof of concept that the approach applies to a wider range of combinatorial problems.

poster session 3 @ blue 5, poster session 7 @ blue 5, oral session 3 @ blue 5, poster session 3, poster session 7, oral session 3

Deep unsupervised hashing has been appreciated in the regime of image retrieval. However, most prior arts failed to detect the semantic components and their relationships behind the images, which makes them lack discriminative power. To make up the defect, we propose a novel \textbf{Deep Semantic Components Hashing (DSCH)}, which involves a common sense that an image normally contains a bunch of semantic components with homology and co-occurrence relationships. Based on this prior, DSCH regards the semantic components as latent variables under the Expectation-Maximization framework and designs a two-step iterative algorithm with the objective of maximum likelihood of training data. Firstly, DSCH constructs a semantic component structure by uncovering the fine-grained semantics components of images with a Gaussian Mixture Modal~(GMM), where an image is represented as a mixture of multiple components, and the semantics co-occurrence are exploited. Besides, coarse-grained semantics components, are discovered by considering the homology relationships between fine-grained components, and the hierarchy organization is then constructed. Secondly, DSCH makes the images close to their semantic component centers at both fine-grained and coarse-grained levels, and also makes the images share similar semantic components close to each other. Extensive experiments on three benchmark datasets demonstrate that the proposed hierarchical semantic components indeed facilitate the hashing model to achieve superior performance.

poster session 3 @ red 1, poster session 8 @ red 1, oral session 3 @ red 1, poster session 3, poster session 8, oral session 3

3D meshes are widely employed to represent geometry structure of 3D shapes. Due to limitation of scanning sensor precision and other issues, meshes are inevitably affected by noise, which hampers the subsequent applications. Convolultional neural networks (CNNs) achieve great success in image processing tasks, including 2D image denoising, and have been proven to own the capacity of modeling complex features at different scales, which is also particularly useful for mesh denoising. However, due to the nature of irregular structure, CNNs-based denosing strategies cannot be trivially applied for meshes. To circumvent this limitation, in the paper, we propose the local surface descriptor (LSD), which is able to transform the local deformable surface around a face into 2D grid representation and thus facilitates the deployment of CNNs to generate denoised face normals. To verify the superiority of LSD, we directly feed LSD into the classical Resnet without any complicated network design. The extensive experimental results show that, compared to the state-of-the-arts, our method achieves encouraging performance with respect to both objective and subjective evaluations.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Deep learning-based approaches have shown remarkable performance in the 3D object detection task. However, they suffer from a catastrophic performance drop on the originally trained classes when incrementally learning new classes without revisiting the old data. This "catastrophic forgetting" phenomenon impedes the deployment of 3D object detection approaches in real-world scenarios, where continuous learning systems are needed. In this paper, we study the unexplored yet important class-incremental 3D object detection problem and present the first solution - SDCoT, a novel static-dynamic co-teaching method. Our SDCoT alleviates the catastrophic forgetting of old classes via a static teacher, which provides pseudo annotations for old classes in the new samples and regularizes the current model by extracting previous knowledge with a distillation loss. At the same time, SDCoT consistently learns the underlying knowledge from new data via a dynamic teacher. We conduct extensive experiments on two benchmark datasets and demonstrate the superior performance of our SDCoT over baseline approaches in several incremental learning scenarios.

poster session 3 @ red 2, poster session 8 @ red 2, oral session 3 @ red 2, poster session 3, poster session 8, oral session 3

Data-free knowledge distillation (DFKD) has recently been attracting increasing attention from research communities, attributed to its capability of compressing a model only using synthetic data. Despite the encouraging results achieved, state-of-the-art DFKD methods still suffer from the inefficiency of data synthesis, making the data-free training process extremely time-consuming and thus inapplicable for large-scale tasks. In this work, we introduce an efficacious scheme, termed as FastDFKD, that allows us to accelerate DFKD by a factor of orders of magnitude. At the heart of our approach is a novel strategy to reuse the shared common features in training data so as to synthesize different data instances. Unlike prior methods that optimize a set of data independently, we propose to learn a meta-synthesizer that seeks common features as the initialization for the fast data synthesis. As a result, FastDFKD achieves data synthesis within only a few steps, significantly enhancing the efficiency of data-free training. Experiments over CIFAR, NYUv2 and ImageNet demonstrate that the proposed FastDFKD achieves 10$\times$ and even 100$\times$ acceleration while preserving performances on par with the state of the art.

poster session 3 @ blue 3, poster session 8 @ blue 3, poster session 3, poster session 8

The task of summarization often requires a non-trivial understanding of the given text at the semantic level. In this work, we essentially incorporate the constituent structure into the single document summarization via the Graph Neural Networks to learn the semantic meaning of tokens. More specifically, we propose a novel hierarchical heterogeneous graph attention network over constituency-based parse trees for syntax-aware summarization. This approach reflects psychological findings that humans will pinpoint specific selection patterns to construct summaries hierarchically. Extensive experiments demonstrate that our model is effective for both the abstractive and extractive summarization tasks on six benchmark datasets from various domains. Moreover, further performance improvement can be obtained by virtue of state-of-the-art pre-trained models. Lastly, we test our model for the code summarization task for border impact.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Accurate 3D object detection from point clouds has become a crucial component in autonomous driving. However, the volumetric representations and the projection methods in previous works fail to establish the relationships between the local point sets. In this paper, we propose Sparse Voxel-Graph Attention Network (SVGA-Net), a novel end-to-end trainable network which mainly contains voxel-graph module and sparse-to-dense regression module to achieve comparable 3D detection tasks from raw LIDAR data. Specifically, SVGA-Net constructs the local complete graph within each divided 3D spherical voxel and global KNN graph through all voxels. The local and global graphs serve as the attention mechanism to enhance the extracted features. In addition, the novel sparse-to-dense regression module enhances the 3D box estimation accuracy through feature maps aggregation at different levels. Experiments on KITTI detection benchmark and Waymo Open dataset demonstrate the efficiency of extending the graph representation to 3D object detection and the proposed SVGA-Net can achieve decent detection accuracy.

poster session 3 @ red 2, poster session 8 @ red 2, poster session 3, poster session 8

Employing additional prior knowledge to model local features as a final fine-grained object representation has become a trend for fine-grained object retrieval (FGOR). A potential limitation of these methods is that they only focus on common parts across the dataset (e.g. head, body or even leg) by introducing additional prior knowledge, but the retrieval of a fine-grained object may rely on category-specific nuances that contribute to category prediction. To handle this limitation, we propose an end-to-end Category-specific Nuance Exploration Network (CNENet) that elaborately discovers category-specific nuances that contribute to category prediction, and semantically aligns these nuances grouped by subcategory without any additional prior knowledge, to directly emphasize the discrepancy among subcategories. Specifically, we design a Nuance Modelling Module that adaptively predicts a group of category-specific response (CARE) maps via implicitly digging into category-specific nuances, specifying the locations and scales for category-specific nuances. Upon this, two nuance regularizations are proposed: 1) semantic discrete loss that forces each CARE map to attend to different spatial regions to capture diverse nuances; 2) semantic alignment loss that constructs a consistent semantic correspondence for each CARE map of the same order with the same subcategory via guaranteeing each instance and its transformed counterpart to be spatially aligned. Moreover, we propose a Nuance Expansion Module, which exploits context appearance information of discovered nuances and refines the prediction of current nuance by its similar neighbors, leading to further improvement on nuance consistency and completeness. Extensive experiments validate that our CNENet consistently yields the best performance under the same settings against most competitive approaches on CUB Birds, Stanford Cars, and FGVC Aircraft datasets.

poster session 3 @ red 1, poster session 8 @ red 1, poster session 3, poster session 8

My doctoral research focuses on understanding semantic knowledge in neural network models trained solely to predict natural language (referred to as language models, or LMs), by drawing on insights from the study of concepts and categories grounded in cognitive science. I propose a framework inspired by 'inductive reasoning,' a phenomenon that sheds light on how humans utilize background knowledge to make inductive leaps and generalize from new pieces of information about concepts and their properties. Drawing from experiments that study inductive reasoning, I propose to analyze semantic inductive generalization in LMs using phenomena observed in human-induction literature, investigate inductive behavior on tasks such as implicit reasoning and emergent feature recognition, and analyze and relate induction dynamics to the learned conceptual representation space.

poster session 3 @ red 5, poster session 1 @ red 5, poster session 3, poster session 1

Several recent works have been dedicated to unsupervised reinforcement learning in a single environment, in which a policy is first pre-trained with unsupervised interactions, and then fine-tuned towards the optimal policy for several downstream supervised tasks defined over the same environment. Along this line, we address the problem of unsupervised reinforcement learning in a class of multiple environments, in which the policy is pre-trained with interactions from the whole class, and then fine-tuned for several tasks in any environment of the class. Notably, the problem is inherently multi-objective as we can trade off the pre-training objective between environments in many ways. In this work, we foster an exploration strategy that is sensitive to the most adverse cases within the class. Hence, we cast the exploration problem as the maximization of the mean of a critical percentile of the state visitation entropy induced by the exploration strategy over the class of environments. Then, we present a policy gradient algorithm, $\alpha$MEPOL, to optimize the introduced objective through mediated interactions with the class. Finally, we empirically demonstrate the ability of the algorithm in learning to explore challenging classes of continuous environments and we show that reinforcement learning greatly benefits from the pre-trained exploration strategy w.r.t. learning from scratch.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

Model generalization to the unseen scenes is crucial to real-world applications, such as autonomous driving, which requires robust vision systems. To enhance the model generalization, domain generalization through learning the domain-invariant representation has been widely studied. However, most existing works learn the shared feature space within multi-source domains but ignore the characteristic of the feature itself (e.g., the feature sensitivity to the domain-specific style). Therefore, we propose the Domain-invariant Representation Learning (DIRL) for domain generalization which utilizes the feature sensitivity as the feature prior to guide the enhancement of the model generalization capability. The guidance reflects in two folds: 1) Feature re-calibration that introduces the Prior Guided Attention Module (PGAM) to emphasize the insensitive features and suppress the sensitive features. 2): Feature whiting that proposes the Guided Feature Whiting (GFW) to remove the feature correlations which are sensitive to the domain-specific style. We construct the domain-invariant representation which suppresses the effect of the domain-specific style on the quality and correlation of the features. As a result, our method is simple yet effective, and can enhance the robustness of various backbone networks with little computational cost. Extensive experiments over multiple domains generalizable segmentation tasks show the superiority of our approach to other methods.

poster session 3 @ red 3, poster session 7 @ red 3, oral session 7 @ red 3, poster session 3, poster session 7, oral session 7

Despite the recent success of end-to-end deep neural networks, there are growing concerns about their lack of logical reasoning abilities, especially on cognitive tasks with perception and reasoning processes. A solution is the neural symbolic learning (NeSyL) method that can effectively utilize pre-defined logic rules to constrain the neural architecture making it perform better on cognitive tasks. However, it is challenging to apply NeSyL to these cognitive tasks because of the lack of supervision, the non-differentiable manner of the symbolic system, and the difficulty to probabilistically constrain the neural network. In this paper, we propose WS-NeSyL, a weakly supervised neural symbolic learning model for cognitive tasks with logical reasoning. First, WS-NeSyL employs a novel back search algorithm to sample the possible reasoning process through logic rules. This sampled process can supervise the neural network as the pseudo label. Based on this algorithm, we can backpropagate gradients to the neural network of WS-NeSyL in a weakly supervised manner. Second, we introduce a probabilistic logic regularization into WS-NeSyL to help the neural network learn probabilistic logic. To evaluate WS-NeSyL, we have conducted experiments on three cognitive datasets, including temporal reasoning, handwritten formula recognition, and relational reasoning datasets. Experimental results show that WS-NeSyL not only outperforms the end-to-end neural model but also beats the state-of-the-art neural symbolic learning models.

poster session 3 @ blue 5, poster session 7 @ blue 5, poster session 3, poster session 7

Amid data privacy concerns, Federated Learning (FL) has emerged as a promising machine learning paradigm that enables privacy-preserving collaborative model training. However, there exists a need for a platform that matches data owners (supply) with model requesters (demand). In this paper, we present CrowdFL, a platform to facilitate the crowdsourcing of FL model training. It coordinates client selection, model training, and reputation management, which are essential steps for the FL crowdsourcing operations. By implementing model training on actual mobile devices, we demonstrate that the platform improves model performance and training efficiency. To the best of our knowledge, it is the first platform to support crowdsourcing-based FL on edge devices.

poster session 3 @ blue 6, poster session 7 @ blue 6, poster session 3, poster session 7

Recovering a dense depth image from sparse LiDAR scans is a challenging task. Despite the popularity of color-guided methods for sparse-to-dense depth completion, they treated pixels equally during optimization, ignoring the uneven distribution characteristics in the sparse depth map and the accumulated outliers in the synthesized ground truth. In this work, we introduce uncertainty-driven loss functions to improve the robustness of depth completion and handle the uncertainty in depth completion. Specifically, we propose an explicit uncertainty formulation for robust depth completion with Jeffrey's prior. A parametric uncertain-driven loss is introduced and translated to new loss functions that are robust to noisy or missing data. Meanwhile, we propose a multiscale joint prediction model that can simultaneously predict depth and uncertainty maps. The estimated uncertainty map is also used to perform adaptive prediction on the pixels with high uncertainty, leading to a residual map for refining the completion results. Our method has been tested on KITTI Depth Completion Benchmark and achieved the state-of-the-art robustness performance in terms of MAE, IMAE, and IRMSE metrics.

poster session 3 @ red 2, poster session 8 @ red 2, oral session 3 @ red 2, poster session 3, poster session 8, oral session 3

Video-based person re-identification~(re-ID) is an important technique in visual surveillance systems which aims to match video snippets of people captured by different cameras. Existing methods are mostly based on convolutional neural networks~(CNNs), whose building blocks either process local neighbor pixels at a time, or, when 3D convolutions are used to model temporal information, suffer from the misalignment problem caused by person movement. In this paper, we propose to overcome the limitations of normal convolutions with a human-oriented graph method. Specifically, features located at person joint keypoints are extracted and connected as a spatial-temporal graph. These keypoint features are then updated by message passing from their connected nodes with a graph convolutional network~(GCN). During training, the GCN can be attached to any CNN-based person re-ID model to assist representation learning on feature maps, whilst it can be dropped after training for better inference speed. Our method brings significant improvements over the CNN-based baseline model on the MARS dataset with generated person keypoints and a newly annotated dataset: PoseTrackReID. It also defines a new state-of-the-art method in terms of top-1 accuracy and mean average precision in comparison to prior works.

poster session 3 @ red 3, poster session 7 @ red 3, poster session 3, poster session 7

An artificial neural network can be trained by uniformly broadcasting a reward signal to units that implement a REINFORCE learning rule. Though this presents a biologically plausible alternative to backpropagation in training a network, the high variance associated with it renders it impractical to train deep networks. The high variance arises from the inefficient structural credit assignment since a single reward signal is used to evaluate the collective action of all units. To facilitate structural credit assignment, we propose replacing the reward signal to hidden units with the change in the $L^2$ norm of the unit's outgoing weight. As such, each hidden unit in the network is trying to maximize the norm of its outgoing weight instead of the global reward, and thus we call this learning method \emph{Weight Maximization}. We prove that Weight Maximization is approximately following the gradient of rewards in expectation. In contrast to backpropagation, Weight Maximization can be used to train both continuous-valued and discrete-valued units. Moreover, Weight Maximization solves several major issues of backpropagation relating to biological plausibility. Our experiments show that a network trained with Weight Maximization can learn significantly faster than REINFORCE and slightly slower than backpropagation. Weight Maximization illustrates an example of cooperative behavior automatically arising from a population of self-interested agents in a competitive game without any central coordination.

poster session 3 @ blue 1, poster session 7 @ blue 1, poster session 3, poster session 7

State of the art neural methods for open information extraction (OpenIE) usually extract triplets (or tuples) iteratively in an autoregressive or predicate-based manner in order not to produce duplicates. In this work, we propose a different approach to the problem that can be equally or more successful. Namely, we present a novel single-pass method for OpenIE inspired by object detection algorithms from computer vision. We use an order-agnostic loss based on bipartite matching that forces unique predictions and a Transformer-based encoder-only architecture for sequence labeling. The proposed approach is faster and shows superior or similar performance in comparison with state of the art models on standard benchmarks in terms of both quality metrics and inference time. Our model sets the new state of the art performance of 67.7% F1 on CaRB evaluated as OIE2016 while being 3.35x faster at inference than previous state of the art. We also evaluate the multilingual version of~our model in~the zero-shot setting for two languages and introduce a~strategy for generating synthetic multilingual data to fine-tune the model for each specific language. In this setting, we show performance improvement of 15\% on multilingual Re-OIE2016, reaching 75% F1 for both Portuguese and Spanish languages. Code and models are available at https://github.com/sberbank-ai/DetIE.

poster session 3 @ red 4, poster session 12 @ red 4, poster session 3, poster session 12

Graph neural networks (GNNs), has been widely used for supervised learning tasks in graphs reaching state-of-the-art results. However, little work was dedicated to creating unbiased GNNs, i.e., where the classification is uncorrelated with sensitive attributes, such as race or gender. Some ignore the sensitive attributes or optimize for the criteria of statistical parity for fairness. However, it has been shown that neither approaches ensure fairness, but rather cripple the utility of the prediction task. In this work, we present a GNN framework that allows optimizing representations for the notion of Equalized Odds fairness criteria. The architecture is composed of three components: (1) a GNN classifier predicting the utility class, (2) a sampler learning the distribution of the sensitive attributes of the nodes given their labels. It generates samples fed into a (3) discriminator that discriminates between true and sampled sensitive attributes using a novel ``permutation loss'' function. Using these components, we train a model to neglect information regarding the sensitive attribute only with respect to its label. To the best of our knowledge, we are the first to optimize GNNs for the equalized odds criteria. We evaluate our classifier over several graph datasets and sensitive attributes and show our algorithm reaches state-of-the-art results.

poster session 3 @ red 3, poster session 7 @ red 3, oral session 3 @ red 3, poster session 3, poster session 7, oral session 3

Recent advances in large-scale pre-training such as GPT-3 allow seemingly high quality text to be generated from a given prompt. However, such generation systems often suffer from problems of hallucinated facts, and are not inherently designed to incorporate useful external information. Grounded generation models appear to offer remedies, but their training typically relies on rarely-available parallel data where corresponding information-relevant documents are provided for context. We propose a framework that alleviates this data constraint by jointly training a grounded generator and document retriever on the language model signal. The model learns to reward retrieval of the documents with the highest utility in generation, and attentively combines them using a Mixture-of-Experts (MoE) ensemble to generate follow-on text. We demonstrate that both generator and retriever can take advantage of this joint training and work synergistically to produce more informative and relevant text in both prose and dialogue generation.

poster session 3 @ red 5, poster session 8 @ red 5, poster session 3, poster session 8

Pre-training language models (LMs) on large-scale unlabeled text data makes the model much easier to achieve exceptional downstream performance than their counterparts directly trained on the downstream tasks. In this work, we study what specific traits in the pre-training data, other than the semantics, make a pre-trained LM superior to their counterparts trained from scratch on downstream tasks. We propose to use artificially constructed datasets as the pre-training data to exclude the effect of semantics, and further control what characteristics the pre-training corpora have. By fine-tuning the pre-trained models on GLUE benchmark, we can learn how beneficial it is to transfer the knowledge from the model trained on the dataset possessing that specific trait. We define and discuss three different characteristics in the artificial dataset: 1) matching the token's uni-gram or bi-gram distribution between pre-training and downstream fine-tuning, 2) the presence of the explicit dependencies among the tokens in a sequence, 3) the length of the implicit dependencies among the tokens in a sequence. Our experiments show that the explicit dependencies in the sequences of the pre-training data are critical to the downstream performance. Our results also reveal that models achieve better downstream performance when pre-trained on a dataset with a longer range of implicit dependencies. Based on our analysis, we find that models pre-trained with artificial datasets are prone to learn spurious correlation in downstream tasks. Our work reveals that even if the LMs are not pre-trained on natural language, they still gain transferability on certain human language downstream tasks once the LMs learn to model the token dependencies in the sequences. This result helps us understand the exceptional transferability of pre-trained LMs.

poster session 3 @ red 4, poster session 12 @ red 4, oral session 12 @ red 4, poster session 3, poster session 12, oral session 12