Papers

The quadratic memory complexity of Transformers prevents long document summarization in low computational resource scenarios. State-of-the-art models need to apply input truncation, thus discarding and ignoring potential summary-relevant contents, leading to a performance drop. Furthermore, such loss is generally destructive for semantic text analytics in the legal domain. In this paper, we propose a novel semantic self-segmentation (Se3) approach for long document summarization to address the critical problems of low-resource regimes, namely to process longer inputs than the GPU memory capacity and produce accurate summaries despite the availability of only a few dozens of training instances. Se3 segments a long input into semantically coherent chunks, allowing Transformers to summarize very long documents without truncation by summarizing each chunk and concatenating the results. Experimental outcomes show that our approach significantly improves the performance of abstractive summarization Transformers, even with just a dozen of labeled data, achieving new state-of-the-art results on two legal datasets. Finally, we perform ablation studies to assess how the different components of our method contribute to the performance gain.

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

Weakly supervised Video Anomaly Detection (VAD) using Multi-Instance Learning (MIL) is usually based on the fact that the anomaly score of an abnormal snippet is higher than that of a normal snippet. In the beginning of training, due to the limited accuracy of the model, it is easy to select the wrong abnormal snippet. In order to reduce the probability of selection errors, we first propose a Multi-Sequence Learning (MSL) method and a hinge-based MSL ranking loss that uses a sequence composed of multiple snippets as an optimization unit. We then design a Transformer-based MSL network to learn both video-level anomaly probability and snippet-level anomaly scores. In the inference stage, we propose to use the video-level anomaly probability to suppress the fluctuation of snippet-level anomaly scores. Finally, since VAD needs to predict the snippet-level anomaly scores, by gradually reducing the length of selected sequence, we propose a self-training strategy to gradually refine the anomaly scores. Experimental results show that our method achieves significant improvements on ShanghaiTech, UCF-Crime, and XD-Violence.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Constrained optimization problems arise frequently in classical machine learning. There exist frameworks addressing constrained optimization, for instance, CVXPY, but so far, in contrast to deep learning frameworks, GPU support is limited. Here, we provide a framework for solving constrained optimization problems on the GPU. The framework allows the user to specify constrained optimization problems in an easy-to-read modeling language. A solver is then automatically generated from this specification. When run on the GPU, the solver outperforms state-of-the-art approaches like CVXPY combined with a GPU-accelerated solver such as cuOSQP or SCS by several orders of magnitude.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

Spatio-temporal representation learning is critical for video self-supervised representation. Recent approaches mainly use contrastive learning and pretext tasks. However, these approaches learn representation by discriminating sampled instances via feature similarity in the latent space while ignoring the intermediate state of the learned representations, which limits the overall performance. In this work, taking into account the degree of similarity of sampled instances as the intermediate state, we propose a novel pretext task - spatio-temporal overlap rate (STOR) prediction. It stems from the observation that humans are capable of discriminating the overlap rates of videos in space and time. This task encourages the model to discriminate the STOR of two generated samples to learn the representations. Moreover, we employ a joint optimization combining pretext tasks with contrastive learning to further enhance the spatio-temporal representation learning. We also study the mutual influence of each component in the proposed scheme. Extensive experiments demonstrate that our proposed STOR task can favor both contrastive learning and pretext tasks and the joint optimization scheme can significantly improve the spatio-temporal representation in video understanding. The code is available at https://github.com/Katou2/CSTP.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Meta reinforcement learning (meta-RL) aims to learn a policy solving a set of training tasks simultaneously and quickly adapting to new tasks. It requires massive amounts of data drawn from training tasks to infer the common structure shared among tasks. Without heavy reward engineering, the sparse rewards in long-horizon tasks exacerbate the problem of sample efficiency in meta-RL. Another challenge in meta-RL is the discrepancy of difficulty level among tasks, which might cause one easy task dominating learning of the shared policy and thus preclude policy adaptation to new tasks. This work introduces a novel objective function to learn an action translator among training tasks. We theoretically verify that the value of the transferred policy with the action translator can be close to the value of the source policy and our objective function (approximately) upper bounds the value difference. We propose to combine the action translator with context-based meta-RL algorithms for better data collection and moreefficient exploration during meta-training. Our approach em-pirically improves the sample efficiency and performance ofmeta-RL algorithms on sparse-reward tasks.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

In this work, we study the maximin share fairness notion for allocation of indivisible goods in the subadditive and fractionally subadditive settings. While previous work refutes the possibility of obtaining an allocation which is better than $1/2$-\textsf{MMS}, the only positive result for the subadditive setting states that when the number of items is equal to $m$, there always exists an $\Omega(1/\log m)$-\textsf{MMS} allocation. Since the number of items may be larger than the number of agents ($n$), such a bound can only imply a weak bound of $\Omega(\frac{1}{n \log n})$-\textsf{MMS} allocation in general. In this work, we improve this gap exponentially to an $\Omega(\frac{1}{\log n \log \log n})$-\textsf{MMS} guarantee. In addition to this, we prove that when the valuation functions are fractionally subadditive, a $1/4.6$-\textsf{MMS} allocation is guaranteed to exist. This also improves upon the previous bound of $1/5$-\textsf{MMS} guarantee for the fractionally subadditive setting.

poster session 1 @ blue 6, poster session 12 @ blue 6, poster session 1, poster session 12

It is well known that natural language models are vulnerable to adversarial attacks, which are mostly input-specific in nature. Recently, it has been shown that there also exist input-agnostic attacks in NLP models, called universal adversarial triggers. However, existing methods to craft universal triggersare data intensive. They require large amounts of data samplesto generate adversarial triggers, which are typically inaccessible by attackers. For instance, previous works take 3000 data samplesper class for the SNLI dataset to generate adversarial triggers. Inthis paper, we present a novel data-free approach, MINIMAL, tomine input-agnostic adversarial triggers from models. Using the triggers produced with our data-free algorithm, we reduce theaccuracy of Stanford Sentiment Treebank’s positive class from 93.6% to 9.6%. Similarly, for the Stanford Natural LanguageInference (SNLI), our single-word trigger reduces the accuracyof the entailment class from 90.95% to less than 0.6%. Despitebeing completely data-free, we get equivalent accuracy drops as data-dependent methods

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

The development of detection methods for oriented object detection remains a challenging task. A considerable obstacle is the wide variation in the shape (e.g., aspect ratio) of objects. Sample selection in general object detection has been widely studied as it plays a crucial role in the performance of the detection method and has achieved great progress. However, existing sample selection strategies still overlook some issues: (1) most of them ignore the object shape information; (2) they do not make a potential distinction between selected positive samples; and (3) some of them can only be applied to either anchor-free or anchor-based methods and cannot be used for both of them simultaneously. In this paper, we propose novel flexible shape-adaptive selection (SA-S) and shape-adaptive measurement (SA-M) strategies for oriented object detection, which comprise an SA-S strategy for sample selection and SA-M strategy for the quality estimation of positive samples. Specifically, the SA-S strategy dynamically selects samples according to the shape information and characteristics distribution of objects. The SA-M strategy measures the localization potential and adds quality information on the selected positive samples. The experimental results on both anchor-free and anchor-based baselines and four publicly available oriented datasets (DOTA, HRSC2016, UCAS-AOD, and ICDAR2015) demonstrate the effectiveness of the proposed method.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Implementations of artificial intelligence (AI) based on deep learning (DL) have proven to be highly successful in many domains, from biomedical imaging to natural language processing, but are still rarely applied in the space industry, particularly for onboard learning of planetary surfaces. In this project, we discuss the utility and limitations of DL, enhanced with topological footprints of the sensed objects, for multi-class classification of planetary surface patterns, in conjunction with tactile and embedded sensing in rover exploratory missions. We consider a Topological Convolutional Network (TCN) model with a persistence-based attention mechanism for supervised classification of various landforms. We study TCN's performance on the Barefoot surface pattern dataset, a novel surface pressure dataset from a prototype tactile rover wheel, known as the Barefoot Rover tactile wheel. Multi-class pattern recognition in the Barefoot data has neither been ever tackled before with DL nor assessed with topological methods. We provide insights into advantages and restrictions of topological DL as the early-stage concept for onboard learning and planetary exploration.

poster session 1 @ red 6, poster session 9 @ red 6, poster session 1, poster session 9

This abstract proposes an approach towards goal-oriented modeling of the detection and modeling complex social phenomena in multiparty discourse in an online political strategy game. We developed a two-tier approach that first encodes sociolinguistic behavior as linguistic features then use reinforcement learning to estimate the advantage afforded to any player. In the first tier, sociolinguistic behavior, such as Friendship and Reasoning, that speakers use to influence others are encoded as linguistic features to identify the persuasive strategies applied by each player in simultaneous two-party dialogues. In the second tier, a reinforcement learning approach is used to estimate a graph-aware reward function to quantify the advantage afforded to each player based on their standing in this multiparty setup. We apply this technique to the game Diplomacy, using a dataset comprising of over 15,000 messages exchanged between 78 users. Our graph-aware approach shows robust performance compared to a context-agnostic setup.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Conceptual representations of meaning have long been the general focus of Artificial Intelligence (AI) towards the fundamental goal of machine understanding, with innumerable efforts made in Knowledge Representation, Speech and Natural Language Processing, Computer Vision, inter alia. Even today, at the core of Natural Language Understanding lies the task of Semantic Parsing, the objective of which is to convert natural sentences into machine-readable representations. Through this paper, we aim to revamp the historical dream of AI, by putting forward a novel, all-embracing, fully semantic meaning representation, that goes beyond the many existing formalisms. Indeed, we tackle their key limits by fully abstracting text into meaning and introducing language-independent concepts and semantic relations, in order to obtain an interlingual representation. Our proposal aims to overcome the language barrier, and connect not only texts across languages, but also images, videos, speech and sound, and logical formulas, across many fields of AI.

poster session 1 @ red 5, poster session 10 @ red 5, oral session 10 @ red 5, poster session 1, poster session 10, oral session 10

Lexical Semantic Change (LSC) detection, also known as Semantic Shift, is the process of identifying and characterizing variations in language usage across different scenarios such as time and domain. It allows us to track the evolution of word senses, as well as to understand the difference between the language used in distinct communities. LSC detection is often done by applying a distance measure over vectors of two aligned word embedding matrices. In this demonstration, we present SenSE, an interactive semantic shift exploration toolkit that provides visualization and explanation of lexical semantic change for an input pair of text sources. Our system focuses on showing how the different alignment strategies may affect the output of an LSC model as well as on explaining semantic change based on the neighbors of a chosen target word, while also extracting examples of sentences where these semantic deviations appear. The system runs as a web application (available at http://sense.mgruppi.me), allowing the audience to interact by configuring the alignment strategies while visualizing the results in a web browser.

poster session 1 @ red 5, poster session 10 @ red 5, poster session 1, poster session 10

$k$-means and $k$-median clustering are powerful unsupervised machine learning techniques. However, due to complicated dependences on all the features, it is challenging to interpret the resulting cluster assignments. Moshkovitz, Dasgupta, Rashtchian, and Frost [ICML 2020] proposed an elegant model of explainable $k$-means and $k$-median clustering. In this model, a decision tree with $k$ leaves provides a straightforward characterization of the data set into clusters. We study two natural algorithmic questions about explainable clustering. (1) For a given clustering, how to find the ``best explanation'' by using a decision tree with $k$ leaves? (2) For a given set of points, how to find a decision tree with $k$ leaves minimizing the $k$-means/median objective of the resulting explainable clustering? To address the first question, we introduce a new model of explainable clustering. Our model, inspired by the notion of outliers in robust statistics, is the following. We are seeking a small number of points (outliers) whose removal makes the existing clustering well-explainable. For addressing the second question, we initiate the study of the model of Moshkovitz et al. from the perspective of multivariate complexity. Our rigorous algorithmic analysis sheds some light on the influence of parameters like the input size, dimension of the data, the number of outliers, the number of clusters, and the approximation ratio, on the computational complexity of explainable clustering.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

Unsupervised graph-level representation learning plays a crucial role in a variety of tasks such as molecular property prediction and community analysis, especially when data annotation is expensive. Currently, most of the best-performing graph embedding methods are based on Infomax principle. The performance of these methods highly depends on the selection of negative samples and hurt the performance, if the samples were not carefully selected. Inter-graph similarity-based methods also suffer if the selected set of graphs for similarity matching is low in quality. To address this, we focus only on utilizing the current input graph for embedding learning. We are motivated by an observation from real-world graph generation processes where the graphs are formed based on one or more global factors which are common to all elements of the graph (e.g., topic of a discussion thread, solubility level of a molecule). We hypothesize extracting these common factors could be highly beneficial. Hence, this work proposes a new principle for unsupervised graph representation learning: Graph-wise Common latent Factor EXtraction (GCFX). We further propose a deep model for GCFX, deepGCFX, based on the idea of reversing the above-mentioned graph generation process which could explicitly extract common latent factors from an input graph and achieve improved results on downstream tasks to the current state-of-the-art. Through extensive experiments and analysis, we demonstrate that, while extracting common latent factors is beneficial for graph-level tasks to alleviate distractions caused by local variations of individual nodes or local neighbourhoods, it also benefits node-level tasks by enabling long-range node dependencies, especially for disassortative graphs.

poster session 1 @ blue 1, poster session 8 @ blue 1, oral session 8 @ blue 1, poster session 1, poster session 8, oral session 8

Knowledge distillation (KD) has recently emerged as a powerful strategy to transfer knowledge from a pre-trained teacher model to a lightweight student, and has demonstrated its unprecedented success over a wide spectrum of applications. In spite of the encouraging results, the KD process \emph{per se} poses a potential threat to network ownership protection, since the knowledge contained in network can be effortlessly distilled and hence exposed to a malicious user. In this paper, we propose a novel framework, termed as Safe Distillation Box~(SDB), that allows us to wrap a pre-trained model in a virtual box for intellectual property protection. Specifically, SDB preserves the inference capability of the wrapped model to all users, but precludes KD from unauthorized users. For authorized users, on the other hand, SDB carries out a knowledge augmentation scheme to strengthen the KD performances and the results of the student model. In other words, all users may employ a model in SDB for inference, but only authorized users get access to KD from the model. The proposed SDB imposes no constraints over the model architecture, and may readily serve as a plug-and-play solution to protect the ownership of a pre-trained network. Experiments across various datasets and architectures demonstrate that, with SDB, the performance of an unauthorized KD drops significantly while that of an authorized gets enhanced, demonstrating the effectiveness of SDB.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

This paper proposes a representational model for image pairs such as consecutive video frames that are related by local pixel displacements, in the hope that the model may shed light on motion perception in primary visual cortex (V1). The model couples the following two components: (1) The vector representations of local contents of images and (2) The matrix representations of local pixel displacements caused by the relative motions between the agent and the objects in the 3D scene. When the image frame undergoes changes due to local pixel displacements, the vectors are multiplied by the matrices that represent the local displacements. Our experiments show that our model can learn Gabor-like filter pairs of quadrature phases. The profiles of the learned filters match those of simple cells in Macaque V1. Moreover, we demonstrate that the model can learn to infer local motions in either a supervised or unsupervised manner. With such a simple model, we achieve competitive results on optical flow estimation.

poster session 1 @ red 3, poster session 11 @ red 3, oral session 1 @ red 3, poster session 1, poster session 11, oral session 1

Open Set Recognition (OSR) has been an emerging topic. Besides recognizing predefined classes, the system needs to reject the unknowns. Prototype learning is a potential manner to handle the problem, as its ability to improve intra-class compactness of representations is much needed in discrimination between the known and the unknowns. In this work, we propose a novel Prototype Mining And Learning (PMAL) framework. It has a prototype mining mechanism before the phase of optimizing embedding space, explicitly considering two crucial properties, namely high-quality and diversity of the prototype set. Concretely, a set of high-quality candidates are firstly extracted from training samples based on data uncertainty learning, avoiding the interference from unexpected noise. Considering the multifarious appearance of objects even in a single category, a diversity-based strategy for prototype set filtering is proposed. Accordingly, the embedding space can be better optimized to discriminate therein the predefined classes and between known and unknowns. Extensive experiments verify the two good characteristics (i.e., high-quality and diversity) embraced in prototype mining, and show the remarkable performance of the proposed framework compared to state-of-the-arts.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Multimodal summarization with multimodal output (MSMO) generates a summary with both textual and visual content. Multimodal news report contains heterogeneous contents, which makes MSMO nontrivial. Moreover, it is observed that different modalities of data in the news report correlate hierarchically. Traditional MSMO methods indistinguishably handle different modalities of data by learning a representation for the whole data, which is not directly adaptable to the heterogeneous contents and hierarchical correlation. In this paper, we propose a hierarchical cross-modality semantic correlation learning model (HCSCL) to learn the intra- and inter-modal correlation existing in the multimodal data. HCSCL adopts a graph network to encode the intra-modal correlation. Then, a hierarchical fusion framework is proposed to learn the hierarchical correlation between text and images. Furthermore, we construct a new dataset with relevant image annotation and image object label information to provide the supervision information for the learning procedure. Extensive experiments on the dataset show that HCSCL significantly outperforms the baseline methods in automatic summarization metrics and fine-grained diversity tests.

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

Information retrieval (IR) is essential in search engines and dialogue systems as well as natural language processing tasks such as open-domain question answering. IR serve an important function in the biomedical domain, where content and sources of scientific knowledge may evolve rapidly. Although neural retrievers have surpassed traditional IR approaches such as TF-IDF and BM25 in standard open-domain question answering tasks, they are still found lacking in the biomedical domain. In this paper, we seek to improve information retrieval (IR) using neural retrievers (NR) in the biomedical domain, and achieve this goal using a three-pronged approach. First, to tackle the relative lack of data in the biomedical domain, we propose a template-based question generation method that can be leveraged to train neural retriever models. Second, we develop two novel pre-training tasks that are closely aligned to the downstream task of information retrieval. Third, we introduce the ``Poly-DPR'' model which encodes each context into multiple context vectors. Extensive experiments and analysis on the BioASQ challenge suggest that our proposed method leads to large gains over existing neural approaches and beats BM25 in the small-corpus setting. We show that BM25 and our method can complement each other, and a simple hybrid model leads to further gains in the large corpus setting.

poster session 1 @ red 2, poster session 10 @ red 2, poster session 1, poster session 10

My work studies the problem of generating scenarios to evaluate interaction between humans and robots. I expect these interactions to grow in complexity as robots become more intelligent and enter our daily lives. However, evaluating such interactions only through user studies, which are the de facto evaluation method in human-robot interaction, will quickly become infeasible as the number of possible scenarios grows exponentially with scenario complexity. Therefore, I propose automatically generating scenarios in simulation to explore the diverse possibility space of scenarios to better understand interaction and avoid costly failures in real world settings.

poster session 1 @ blue 6, poster session 2 @ blue 6, poster session 1, poster session 2

Temporal action segmentation classifies the action of each frame in (long) video sequences. Due to the high cost of frame-wise labeling, we propose the first semi-supervised method for temporal action segmentation. Our method hinges on unsupervised representation learning, which, for temporal action segmentation, poses unique challenges. Actions in untrimmed videos vary in length and have unknown labels and start/end times. Ordering of actions across videos may also vary. We propose a novel way to learn frame-wise representations from temporal convolutional networks (TCNs) by clustering input features with added time-proximity condition and multi-resolution similarity. By merging representation learning with conventional supervised learning, we develop an ``Iterative-Contrast-Classify (ICC)'' semi-supervised learning scheme. With more labelled data, ICC progressively improves in performance; ICC semi-supervised learning, with 40% labelled videos, performs similar to fully-supervised counterparts. Our ICC improves MoF by {+1.8, +5.6, +2.5}% on Breakfast, 50Salads and GTEA respectively for 100% labelled videos.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Real world tasks are hierarchical and compositional. Tasks can be composed of multiple subtasks (or sub-goals) that are dependent on each other. These subtasks are defined in terms of entities (e.g., "apple", "pear") that can be recombined to form new subtasks (e.g., "pickup apple", and "pickup pear"). To solve these tasks efficiently, an agent must infer subtask dependencies (e.g. an agent must execute "pickup apple" before "place apple in pot"), and generalize the inferred dependencies to new subtasks (e.g. "place apple in pot" is similar to "place apple in pan"). Moreover, an agent may also need to solve unseen tasks, which can involve unseen entities. To this end, we formulate parameterized subtask graph inference (PSGI), a method for modeling subtask dependencies using first-order logic with factored entities. To facilitate this, we learn parameter attributes in a zero-shot manner, which are used as quantifiers (e.g. is_pickable(X)) for the factored subtask graph. We show this approach accurately learns the latent structure on hierarchical and compositional tasks more efficiently than prior work, and show PSGI can generalize by modelling structure on subtasks unseen during adaptation.

poster session 1 @ blue 1, poster session 8 @ blue 1, oral session 1 @ blue 1, poster session 1, poster session 8, oral session 1

We study fair and efficient allocation of divisible goods, in an online manner, among n agents. The goods arrive online in a sequence of T time periods. The agents' values for a good are revealed only after its arrival, and the online algorithm needs to fractionally allocate the good, immediately and irrevocably, among the agents. Towards a unifying treatment of fairness and economic efficiency objectives, we develop an algorithmic framework for finding online allocations to maximize the generalized mean of the values received by the agents. In particular, working with the assumption that each agent's value for the grand bundle of goods is appropriately scaled, we address online maximization of p-mean welfare. Parameterized by an exponent term p in (-infinity, 1], these means encapsulate a range of welfare functions, including social welfare (p=1), egalitarian welfare (p to -infinity), and Nash social welfare (p to 0). We present a simple algorithmic template that takes a threshold as input and, with judicious choices for this threshold, leads to both universal and tailored competitive guarantees. First, we show that one can compute online a single allocation that O (n^{1/2} log n)-approximates the optimal p-mean welfare for all p<=1. The existence of such a universal allocation is interesting in and of itself. Moreover, this universal guarantee achieves essentially tight competitive ratios for specific values of p. Next, we obtain improved competitive ratios for different ranges of p by executing our algorithm with p-specific thresholds, e.g., we provide O(log^3 n)-competitive ratio for all p in (-1/log 2n, 1). We complement our positive results by establishing lower bounds to show that our guarantees are essentially tight for a wide range of the exponent parameter.

poster session 1 @ blue 6, poster session 12 @ blue 6, oral session 1 @ blue 6, poster session 1, poster session 12, oral session 1

The paper studies strategic abilities that rise from restrictions on the information sharing in multi-agent systems. The main technical result is a sound and complete logical system that describes the interplay between the knowledge and the strategic ability modalities.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

In this paper, we propose Deep amortized Relational Model (DaRM) with group-wise hierarchical generative process for community discovery and link prediction on relational data (e.g., graph, network). It provides an efficient neural relational model architecture by grouping nodes in a group-wise view rather than node-wise or edge-wise view. DaRM simultaneously learns what makes a group, how to divide nodes into groups, and how to adaptively control the number of groups. The dedicated group generative process is able to sufficiently exploit pair-wise or higher-order interactions between data points in both inter-group and intra-group, which is useful to sufficiently mine the hidden structure among data. A series of experiments have been conducted on both synthetic and real-world datasets. The experimental results demonstrated that DaRM can obtain high performance on both community detection and link prediction tasks.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Conditional density estimation (CDE) is the task of estimating the probability of an event conditioned on some inputs. A neural network (NN) can also be used to compute the output distribution for continuous-domain, which can be viewed as an extension of regression task. Nevertheless, it is difficult to explicitly approximate a distribution without knowing the information of its general form a priori. In order to fit an arbitrary conditional distribution, discretizing the continuous domain into bins is an effective strategy, as long as we have sufficiently narrow bins and very large data. However, collecting enough data is often hard to reach and falls far short of that ideal in many circumstances, especially in multivariate CDE for the curse of dimensionality. In this paper, we demonstrate the benefits of modeling free-form conditional distributions using a deconvolution-based neural net framework, coping with data deficiency problems in discretization. It has the advantage of being flexible but also takes advantage of the hierarchical smoothness offered by the deconvolution layers. We compare our method to a number of other density-estimation approaches and show that our Deconvolutional Density Network (DDN) outperforms the competing methods on many univariate and multivariate tasks. The code of DDN is available at https://github.com/NBICLAB/DDN

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

This paper tries to resolve long waiting time to find a matching person in player versus player mode of online sports games, such as baseball, soccer and basketball. In player versus player mode, game playing AI which is instead of player needs to be not just smart as human but also show variety to improve user experience against AI. Therefore a need to design game playing AI agents with diverse personalized styles rises. To this end, we propose a personalized game AI which encodes user style vectors and card style vectors with a general DNN, named UCSM-DNN. Extensive experiments show that UCSM-DNN shows improved performance in terms of personalized styles, which enrich user experiences. UCSM-DNN has already been integrated into popular mobile baseball game: MaguMagu 2021 as personalized game AI.

poster session 1 @ red 5, poster session 10 @ red 5, poster session 1, poster session 10

This paper exploits self-supervised learning (SSL) to learn more accurate and robust representations from the user-item interaction graph. Particularly, we propose a novel SSL model that effectively leverages contrastive multi-view learning and pseudo-siamese network to construct a pre-training and post-training framework. Moreover, we present three graph augmentation techniques during the pre-training stage and explore the effects of combining different augmentations, which allow us to learn general and robust representations for the GNN-based recommendation. Simple experimental evaluations on real-world datasets show that the proposed solution significantly improves the recommendation accuracy, especially for sparse data, and is also noise resistant.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

Video moment localization aims at localizing the video segments which are most related to the given free-form natural language query. The weakly supervised setting, where only video level description is available during training, is getting more and more attention due to its lower annotation cost. Prior weakly supervised methods mainly use sliding windows to generate temporal proposals, which are independent of video content and low quality, and train the model to distinguish matched video-query pairs and unmatched ones collected from different videos, while neglecting what the model needs is to distinguish the unaligned segments within the video. In this work, we propose a novel weakly supervised solution by introducing Contrastive Negative sample Mining (CNM). Specifically, we use a learnable Gaussian mask to generate positive samples, highlighting the video frames most related to the query, and consider other frames of the video and the whole video as easy and hard negative samples respectively. We then train our network with the Intra-Video Contrastive loss to make our positive and negative samples more discriminative. Our method has two advantages: (1) Our proposal generation process with a learnable Gaussian mask is more efficient and makes our positive sample higher quality. (2) The more difficult intra-video negative samples enable our model to distinguish highly confusing scenes. Experiments on two datasets show the effectiveness of our method. Code can be found at https://github.com/minghangz/cnm.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Hierarchical structures are popular in recent vision transformers, however, they require sophisticated designs and massive datasets to work well. In this paper, we explore the idea of nesting basic local transformers on non-overlapping image blocks and aggregating them in a hierarchical way. We find that the block aggregation function plays a critical role in enabling cross-block non-local information communication. This observation leads us to design a simplified architecture that requires minor code changes upon the original vision transformer. The benefits of the proposed judiciously-selected design are threefold: (1)~NesT converges faster and requires much less training data to achieve good generalization on both ImageNet and small datasets like CIFAR; (2)~when extending our key ideas to image generation, NesT leads to a strong decoder that is 8$\times$ faster than previous transformer-based generators; and (3)~we show that decoupling the feature learning and abstraction processes via this nested hierarchy in our design enables constructing a novel method (named GradCAT) for visually interpreting the learned model.

poster session 1 @ red 3, poster session 11 @ red 3, oral session 1 @ red 3, poster session 1, poster session 11, oral session 1

This paper proposes a novel approach to object detection on drone imagery, namely Multi-Proxy Detection Network with Unified Foreground Packing (UFPMP-Det). To deal with the numerous instances of very small scales, different from the common solution that divides the high-resolution input image into quite a number of chips with low foreground ratios to perform detection on them each, the Unified Foreground Packing (UFP) module is designed, where the sub-regions given by a coarse detector are initially merged through clustering to suppress background and the resulting ones are subsequently packed into a mosaic for a single inference, thus significantly reducing overall time cost. Furthermore, to address the more serious confusion between inter-class similarities and intra-class variations of instances, which deteriorates detection performance but is rarely discussed, the Multi-Proxy Detection Network (MP-Det) is presented to model object distributions in a fine-grained manner by employing multiple proxy learning, and the proxies are enforced to be diverse by minimizing a Bag-of-Instance-Words (BoIW) guided optimal transport loss. By such means, UFPMP-Det largely promotes both the detection accuracy and efficiency. Extensive experiments are carried out on the widely used VisDrone and UAVDT datasets, and UFPMP-Det reports new state-of-the-art scores at a much higher speed, highlighting its advantages.

poster session 1 @ red 1, poster session 11 @ red 1, oral session 11 @ red 1, poster session 1, poster session 11, oral session 11

Monaural speech enhancement (SE) at an extremely low signal-to-noise ratio (SNR) condition is a challenging problem and rarely investigated in previous studies. Most SE methods experience failures in this situation due to three major factors: overwhelmed vocals, expanded SNR range, and short-sighted feature processing modules. In this paper, we present a novel and general training paradigm dubbed repetitive learning (RL). Unlike curriculum learning that focuses on learning multiple different tasks sequentially, RL is more inclined to learn the same content repeatedly where the knowledge acquired in previous stages can be used to facilitate calibrating feature representations. We further propose an RL-based end-to-end SE method named SERL. Experimental results on TIMIT dataset validate the superior performance of our method.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Video object detection has been an important yet challenging topic in computer vision. Traditional methods mainly focus on designing the image-level or box-level feature propagation strategies to exploit temporal information. This paper argues that with a more effective and efficient feature propagation framework, video object detectors can gain improvement in terms of both accuracy and speed. For this purpose, this paper studies object-level feature propagation, and proposes an object query propagation (QueryProp) framework for high-performance video object detection. The proposed QueryProp contains two propagation strategies: 1) query propagation is performed from sparse key frames to dense non-key frames to reduce the redundant computation on non-key frames; 2) query propagation is performed from previous key frames to the current key frame to improve feature representation by temporal context modeling. To further facilitate query propagation, an adaptive propagation gate is designed to achieve flexible key frame selection. We conduct extensive experiments on the ImageNet VID dataset. QueryProp achieves comparable accuracy with state-of-the-art methods and strikes a decent accuracy/speed trade-off.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Event camera has offered promising alternative for visual perception, especially in high speed and high dynamic range scenes. Recently, many deep learning methods have shown great success in providing model-free solutions to many event-based problems, such as optical flow estimation. However, existing deep learning methods did not address the importance of temporal information well from the perspective of architecture design and cannot effectively extract spatio-temporal features. Another line of research that utilizes Spiking Neural Network suffers from training issues for deeper architecture. To address these points, a novel input representation is proposed that captures the events temporal distribution for signal enhancement. Moreover, we introduce a spatio-temporal recurrent encoding-decoding neural network architecture for event-based optical flow estimation, which utilizes Convolutional Gated Recurrent Units to extract feature maps from a series of event images. Besides, our architecture allows some traditional frame-based core modules, such as correlation layer and iterative residual refine scheme, to be incorporated. The network is end-to-end trained with self-supervised learning on the Multi-Vehicle Stereo Event Camera dataset. We have shown that it outperforms all the existing state-of-the-art methods by a large margin.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Domain adaptation aims to leverage source domain knowledge to predict target domain labels. Most domain adaptation methods tackle a single-source, single-target scenario, whereas source and target domain data can often be subdivided into data from different distributions in real-life applications (e.g., when the distribution of the collected data changes with time). However, such subdomains are rarely given and should be discovered automatically. To this end, some recent domain adaptation works seek separations of hidden subdomains, w.r.t. a known or fixed number of subdomains. In contrast, this paper introduces a new subdomain combination method that leverages a variable number of subdomains. Precisely, we propose to use an inter-subdomain divergence maximization criterion to exploit hidden subdomains. Besides, our proposition stands in a target-to-source domain adaptation scenario, where one exploits a pre-trained source model as a black box; thus, the proposed method is model-agnostic. By providing interpretability at two complementary levels (transformation and subdomain levels), our method can also be easily interpreted by practitioners with or without machine learning backgrounds. Experimental results over two fraud detection datasets demonstrate the efficiency of our method.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

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

Pedestrian trajectory prediction is crucial in many practical applications due to the diversity of pedestrian movements, such as social interactions and individual motion behaviors. With similar observable trajectories and social environments, different pedestrians may make completely different future decisions. However, most existing methods only focus on the frequent modal of the trajectory and thus are difficult to generalize to the peculiar scenario, which leads to the decline of the multimodal fitting ability when facing similar scenarios. In this paper, we propose a complementary attention gated network (CAGN) for pedestrian trajectory prediction, in which a dual-path architecture including normal and inverse attention is proposed to capture both frequent and peculiar modals in spatial interactions and temporal motion patterns, respectively. Specifically, a complementary block is proposed to guide normal and inverse attention, which are then be summed with learnable weights to get attention features by a gated network. Subsequently, multiple trajectory distributions are estimated based on the fused spatio-temporal attention features due to the multimodality of future trajectory. Experimental results on the popular benchmark datasets, i.e., the ETH, and the UCY, demonstrate that our method outperforms state-of-the-art methods by 13.8% in Average Displacement Error (ADE) and 10.4% in Final Displacement Error (FDE).

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

The goal of a summary is to concisely state the most important information in a document. With this principle in mind, we introduce new reference-free summary evaluation metrics that use a pretrained language model to estimate the information content shared between a document and its summary. These metrics are a modern take on the Shannon Game, a method for summary quality scoring proposed decades ago, where we replace human annotators with language models. We also view these metrics as an extension of BLANC, a recently proposed approach to summary quality measurement based on the performance of a language model with and without the help of a summary. Using transformer based language models, we empirically verify that our metrics achieve state-of-the-art correlation with human judgement of the summary quality dimensions of both coherence and relevance, as well as competitive correlation with human judgement of consistency and fluency.

poster session 1 @ red 4, poster session 8 @ red 4, oral session 1 @ red 4, poster session 1, poster session 8, oral session 1

Spectral graph convolution is extremely time-consuming for large graphs, thus existing Graph Convolutional Networks (GCNs) reconstruct the kernel by a polynomial, which is (almost) fixed. To extract features from the graph data by learning kernels, Low-pass Collaborative Filter Network (LCFN) was proposed as a new paradigm with trainable kernels. However, there are two demerits of LCFN: (1) The hypergraphs in LCFN are constructed by mining 2-hop connections of the user-item bipartite graph, thus 1-hop connections are not used, resulting in serious information loss. (2) LCFN follows the general network structure of GCNs, thus the structure is suboptimal. To address these issues, we utilize the bipartite graph to define the graph space directly and explore the best network structure based on experiments. Comprehensive experiments on two real-world datasets demonstrate the effectiveness of the proposed model.

poster session 1 @ blue 5, poster session 11 @ blue 5, oral session 11 @ blue 5, poster session 1, poster session 11, oral session 11

Pseudo bounding boxes from the self-training paradigm are inevitably noisy for semi-supervised object detection. To cope with that, a dual decoupling training framework is proposed in the present study, i.e. clean and noisy data decoupling, and classifcation and localization task decoupling. In the frst decoupling, two-level thresholds are used to categorize pseudo boxes into three groups, i.e. clean backgrounds, noisy foregrounds and clean foregrounds. With a specially designed noise-bypass head focusing on noisy data, backbone networks can extract coarse but diverse information; and meanwhile, an original head learns from clean samples for more precise predictions. In the second decoupling, we take advantage of the two-head structure for better evaluation of localization quality, thus the category label and location of a pseudo box can remain independent of each other during training. The approach of two-level thresholds is also applied to group pseudo boxes into three sections of different location accuracy. We outperform existing works by a large margin on VOC datasets, reaching 54.8 mAP(+1.8), and even up to 55.9 mAP(+1.5) by leveraging MS-COCO train2017 as extra unlabeled data. On MS-COCO benchmark, our method also achieves about 1.0 mAP improvements averaging across protocols compared with the prior state-of-the-art.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

There have been two streams in the 3D detection from point clouds: single-stage methods and two-stage methods. While the former is more computationally efficient, the latter usually provides better detection accuracy. By carefully examining the two-stage approaches, we have found that if appropriately designed, the first stage can produce accurate box regression. In this scenario, the second stage mainly rescores the boxes such that the boxes with better localization get selected. From this observation, we have devised a single-stage anchor-free network that can fulfill these requirements. This network, named AFDetV2, extends the previous work by incorporating a self-calibrated convolution block in the backbone, a keypoint auxiliary supervision, and an IoU prediction branch in the multi-task head. We take a simple product of the predicted IoU score with the classification heatmap to form the final classification confidence. The enhanced backbone strengthens the box localization capability, and the rescoring approach effectively joins the object presence confidence and the box regression accuracy. As a result, the detection accuracy is drastically boosted in the single-stage. To evaluate our approach, we have conducted extensive experiments on the Waymo Open Dataset and the nuScenes Dataset. We have observed that our AFDetV2 achieves the state-of-the-art results on these two datasets, superior to all the prior arts, including both the single-stage and the two-stage 3D detectors. AFDetV2 won the 1st place in the Real-Time 3D Detection of the Waymo Open Dataset Challenge 2021. In addition, a variant of our model AFDetV2-Base was entitled the "Most Efficient Model" by the Challenge Sponsor, showing a superior computational efficiency. To demonstrate the generality of this single-stage method, we have also applied it to the first stage of the two-stage networks. Without exception, the results show that with the strengthened backbone and the rescoring approach, the second stage refinement is no longer needed.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Vision transformers (ViTs) have been an alternative design paradigm to convolutional neural networks (CNNs). However, the training of ViTs is much harder than CNNs, as it is sensitive to the training parameters, such as learning rate, optimizer and warmup epoch. The reasons for training difficulty are empirically analysed in ~\cite{xiao2021early}, and the authors conjecture that the issue lies with the \textit{patchify-stem} of ViT models and propose that early convolutions help transformers see better. In this paper, we further investigate this problem and extend the above conclusion: only early convolutions do not help for stable training, but the scaled ReLU operation in the \textit{convolutional stem} (\textit{conv-stem}) matters. We verify, both theoretically and empirically, that scaled ReLU in \textit{conv-stem} not only improves training stabilization, but also increases the diversity of patch tokens, thus boosting peak performance with a large margin via adding few parameters and flops. In addition, extensive experiments are conducted to demonstrate that previous ViTs are far from being well trained, further showing that ViTs have great potential to be a better substitute of CNNs.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Unsupervised video anomaly detection, a task that requires no labeled normal/abnormal training data in any form, is challenging yet of great importance to both industrial applications and academic research. Existing methods typically follow an iterative pseudo label generation process. However, they lack a principled analysis of the impact of such pseudo label generation on training. Furthermore, the long-range temporal dependencies also has been overlooked, which is unreasonable since the definition of an abnormal event depends on the long-range temporal context. To this end, first, we propose a causal graph to analyze the confounding effect of the pseudo label generation process. Then, we introduce a simple yet effective causal inference based framework to disentangle the noisy pseudo label's impact. Finally, we perform counterfactual based model ensemble that blends long-range temporal context with local image context in inference to make final anomaly detection. Extensive experiments on six standard benchmark datasets show that our proposed method significantly outperforms previous state-of-the-art methods, demonstrating our framework's effectiveness.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

As a decisive part in the success of Mobility-as-a-Service (MaaS), spatio-temporal predictive modeling for crowd movements is a challenging task particularly considering scenarios where societal events drive mobility behavior deviated from the normality. While tremendous progress has been made to model high-level spatio-temporal regularities with deep learning, most, if not all of the existing methods are neither aware of the dynamic interactions among multiple transport modes nor adaptive to unprecedented volatility brought by potential societal events. In this paper, we are therefore motivated to improve the canonical spatio-temporal network (ST-Net) from two perspectives: (1) design a heterogeneous mobility information network (HMIN) to explicitly represent intermodality in multimodal mobility; (2) propose a memory-augmented dynamic filter generator (MDFG) to generate sequence-specific parameters in an on-the-fly fashion for various scenarios. The enhanced event-aware spatio-temporal network, namely EAST-Net, is evaluated on several real-world datasets with a wide variety and coverage of societal events. Both quantitative and qualitative experimental results verify the superiority of our approach compared with the state-of-the-art baselines. Code and data are published on https://github.com/underdoc-wang/EAST-Net.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

Grounded Situation Recognition (GSR), i.e., recognizing the salient activity (or verb) category in an image (e.g.,buying) and detecting all corresponding semantic roles (e.g.,agent and goods), is an essential step towards “human-like” event understanding. Since each verb is associated with a specific set of semantic roles, all existing GSR methods resort to a two-stage framework: predicting the verb in the first stage and detecting the semantic roles in the second stage. However, there are obvious drawbacks in both stages: 1) The widely-used cross-entropy (XE) loss for object recognition is insufficient in verb classification due to the large intra-class variation and high inter-class similarity among daily activities. 2) All semantic roles are detected in an autoregressive manner, which fails to model the complex semantic relations between different roles. To this end, we propose a novel SituFormerfor GSR which consists of a Coarse-to-Fine Verb Model (CFVM) and a Transformer-based Noun Model (TNM). CFVM is a two-step verb prediction model: a coarse-grained model trained with XE loss first proposes a set of verb candidates, and then a fine-grained model trained with triplet loss re-ranks these candidates with enhanced verb features (not only separable but also discriminative). TNM is a transformer-based semantic role detection model, which detects all roles parallelly. Owing to the global relation modeling ability and flexibility of the transformer decoder, TNM can fully explore the statistical dependency of the roles. Extensive validations on the challenging SWiG benchmark show that SituFormer achieves a new state-of-the-art performance with significant gains under various metrics. Code is available at https://github.com/kellyiss/SituFormer.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Semantic scene completion (SSC) aims at predicting the semantic categories and 3D volumetric occupancies of a scene simultaneously. The RGB-D images have been broadly used for providing the semantic and geometry information of the objects in existing semantic scene completion methods. However, existing works use concatenation, element-wise summation and weighted summation to fuse RGB-D data. These strategies ignore the large discrepancy of RGB-D data and the uncertainty measurements of depth data. To solve this problem, we propose the \emph{Frequency Fusion Network} (FFNet), a novel method boosting semantic scene completion by better utilizing RGB-D data. It first explicitly correlates the RGB-D data in the frequency domain, which is different from the features directly extracted by the convolution operation. Then, the correlation information is used to guide the RGB-assisted depth features and depth-assisted RGB features. Further, considering the properties of different frequency components of RGB-D features, we propose a learnable elliptical mask to decompose the features, and attend to different frequency bands to facilitate the correlation process of RGB-D data. We evaluate FFNet intensively on the public SSC benchmarks, where FFNet surpasses the state-of-the-art methods. Our code will be made publicly available.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Due to high-speed motion blur and challenging illumination, conventional frame-based cameras have encountered an important challenge in object detection tasks. Neuromorphic cameras that output asynchronous visual streams instead of intensity frames, by taking the advantage of high temporal resolution and high dynamic range, have brought a new perspective to address the challenge. In this paper, we propose a novel problem setting, retinomorphic object detection, which is the first trial that integrates foveal-like and peripheral-like visual streams. Technically, we first build a large-scale multimodal neuromorphic object detection dataset (i.e., PKU-Vidar-DVS) over 215.5k spatio-temporal synchronized labels. Then, we design temporal aggregation representations to preserve the spatio-temporal information from asynchronous visual streams. Finally, we present a novel bio-inspired unifying framework to fuse two sensing modalities via a dynamic interaction mechanism. Our experimental evaluation shows that our approach has significant improvements over the state-of-the-art methods with the single-modality, especially in high-speed motion and low-light scenarios. We hope that our work will attract further research into this newly identified, yet crucial research direction. Our dataset can be available at https://www.pkuml.org/resources/pku-vidar-dvs.html.

poster session 1 @ red 3, poster session 11 @ red 3, oral session 11 @ red 3, poster session 1, poster session 11, oral session 11

Transformer models have demonstrated their promising potential and achieved excellent performance on a series of computer vision tasks. However, the huge computational cost of vision transformers hinders their deployment and application to edge devices. Recent works have proposed to find and remove the unimportant units of vision transformers. Despite achieving remarkable results, these methods take one dimension of network width into consideration and ignore network depth, which is another important dimension for pruning vision transformers. Therefore, we propose a Width & Depth Pruning (WDPruning) framework that reduces both width and depth dimensions simultaneously. Specifically, for width pruning, a set of learnable pruning-related parameters is used to adaptively adjust the width of transformer. For depth pruning, we introduce several shallow classifiers by using the intermediate information of the transformer blocks, which allows images to be classified by shallow classifiers instead of the deeper classifiers. In the inference period, all of the blocks after shallow classifiers can be dropped so they don’t bring additional parameters and computation. Experimental results on benchmark datasets demonstrate that the proposed method can significantly reduce the computational costs of mainstream vision transformers such as DeiT and Swin Transformer with a minor accuracy drop. In particular, on ILSVRC-12, we achieve over 22% pruning ratio of FLOPs by compressing DeiT-Base, even with an increase of 0.14% Top-1 accuracy.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

We study the problem of allocating indivisible goods among strategic agents. We focus on settings wherein monetary transfers are not available and each agent's private valuation is a submodular function with binary marginals, i.e., the agents' valuations are matroid-rank functions. In this setup, we establish a notable dichotomy between two of the most well-studied fairness notions in discrete fair division; specifically, between envy-freeness up to one good (EF1) and maximin shares (MMS). First, we show that a Pareto-efficient mechanism of Babaioff et al. (2021) is group strategy-proof for finding EF1 allocations, under matroid-rank valuations. The group strategy-proofness guarantee strengthens the result of Babaioff et al. (2021), that establishes truthfulness (individually for each agent) in the same context. Our result also generalizes a work of Halpern et al. (2020), from binary additive valuations to the matroid-rank case. Next, we establish that an analogous positive result cannot be achieved for MMS, even when considering truthfulness on an individual level. Specifically, we prove that, for matroid-rank valuations, there does not exist a truthful mechanism that is index oblivious, Pareto efficient, and maximin fair. For establishing our results, we develop a characterization of truthful mechanisms for matroid-rank functions. This characterization in fact holds for a broader class of valuations (specifically, holds for binary XOS functions) and might be of independent interest.

poster session 1 @ blue 6, poster session 12 @ blue 6, oral session 1 @ blue 6, poster session 1, poster session 12, oral session 1

We study a dynamic traffic assignment model, where agents base their instantaneous routing decisions on real-time delay predictions. We formulate a mathematically concise model and derive properties of the predictors that ensure a dynamic prediction equilibrium exists. We demonstrate the versatility of our framework by showing that it subsumes the well-known full information and instantaneous information models, in addition to admitting further realistic predictors as special cases. We complement our theoretical analysis by an experimental study, in which we systematically compare the induced average travel times of different predictors, including a machine-learning model trained on data gained from previously computed equilibrium flows, both on a synthetic and a real road network.

poster session 1 @ blue 6, poster session 12 @ blue 6, poster session 1, poster session 12

Most of the existing methods for \textit{search result diversification} (SRD) appeal to the \textit{greedy strategy} for generating diversified results, which is formulated as a sequential process of selecting documents one-by-one, and the locally optimal choice is made at each round. Unfortunately, this strategy suffers from the following shortcomings: (1) Such a one-by-one selection process is rather time-consuming for both training and inference. (2) It works well on the premise that the preceding choices are optimal or close to the optimal solution. (3) The mismatch between the objective function used in training and the final evaluation measure used in testing has not been taken into account. We propose a novel framework through direct metric optimization for SRD (referred to as \textit{MO4SRD}) based on the \textit{score-and-sort strategy}. Specifically, we represent the diversity score of each document that determines its rank position based on a probability distribution. These distributions over scores naturally give rise to expectations over rank positions. Armed with this advantage, we can get the differentiable variants of the widely used diversity metrics. Thanks to this, we are able to directly optimize the evaluation measure used in testing. Moreover, we have devised a novel probabilistic neural scoring function. It jointly scores candidate documents by taking into account both cross-document interaction and permutation equivariance, which makes it possible to generate a diversified ranking via a simple sorting. The experimental results on benchmark collections show that the proposed method achieves significantly improved performance over the state-of-the-art results.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

There are two key issues that limit further improvements in the performance of existing rotational detectors: 1)Periodic sudden change of the parameters in the rotating bounding box (RBBox) definition causes a numerical discontinuity in the loss (such as smoothL1 loss). 2)There is a gap of optimization asynchrony between the loss in the RBBox regression and evaluation metrics. In this paper, we define a new distance formulation between two convex polygons describing the overlapping degree and non-overlapping degree. Based on this smooth distance, we propose a loss called Polygon-to-Polygon distance loss (P2P Loss). The distance is derived from the area sum of triangles specified by the vertexes of one polygon and the edges of the other. Therefore, the P2P Loss is continuous, differentiable, and inherently free from any RBBox definition. Our P2P Loss is not only consistent with the detection metrics but also able to measure how far, as well as how similar, an RBBox is from another one even when they are completely non-overlapping. These features allow the RetinaNet using the P2P Loss to achieve 79.15% mAP on the DOTA dataset, which is quite competitive compared with many state-of-the-art rotated object detectors.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

We propose a new discrete model for simultaneously spreading two opinions within a social network inspired by the famous Target Set Selection problem. We are given a social network, a seed-set of agents for each opinion, and two thresholds per agent. The first threshold represents the willingness of an agent to adopt an opinion if she has no opinion at all, while the second threshold states the readiness to acquire a second opinion. The goal is to add as few agents as possible to the initial seed-sets such that, once the process started with these seed-set stabilises, each agent has either both opinions or none. We perform an initial study of its computational complexity. It is not surprising that the problem is NP-hard even in quite restricted settings. Therefore, we investigate the complexity of the problem from the parameterized point-of-view with special focus on sparse networks, which appears often in practice. Among other things, we show that the proposed problem is in the FPT complexity class if we parameterize by the vertex cover number of the underlying graph.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

Given an initial resource allocation, where some agents may envy others or where a different distribution of resources might lead to higher social welfare, our goal is to improve the allocation without reassigning resources. We consider a sharing concept allowing resources being shared with social network neighbors of the resource owners. To this end, we introduce a formal model that allows a central authority to compute an optimal sharing between neighbors based on an initial allocation. Advocating this point of view, we focus on the most basic scenario where a resource may be shared by two neighbors in a social network and each agent can participate in a bounded number of sharings. We present algorithms for optimizing utilitarian and egalitarian social welfare of allocations and for reducing the number of envious agents. In particular, we examine the computational complexity with respect to several natural parameters. Furthermore, we study cases with restricted social network structures and, among others, devise polynomial-time algorithms in path- and tree-like (hierarchical) social networks.

poster session 1 @ blue 6, poster session 12 @ blue 6, oral session 1 @ blue 6, poster session 1, poster session 12, oral session 1

Transformers have sprung up in the field of computer vision. In this work, we explore whether the core self-attention module in Transformer is the key to achieving excellent performance in image recognition. To this end, we build an attention-free network called sMLPNet based on the existing MLP-based vision models. Specifically, we replace the MLP module in the token-mixing step with a novel sparse MLP (sMLP) module. For 2D image tokens, sMLP applies 1D MLP along the axial directions and the parameters are shared among rows or columns. By sparse connection and weight sharing, sMLP module significantly reduces the number of model parameters and computational complexity, avoiding the common over-fitting problem that plagues the performance of MLP-like models. When only trained on the ImageNet-1K dataset, the proposed sMLPNet achieves 81.9% top-1 accuracy with only 24M parameters, which is much better than most CNNs and vision Transformers under the same model size constraint. When scaling up to 66M parameters, sMLPNet achieves 83.4% top-1 accuracy, which is on par with the state-of-the-art Swin Transformer. The success of sMLPNet suggests that the self-attention mechanism is not necessarily a silver bullet in computer vision. The code and models are publicly available at https://github.com/microsoft/SPACH.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Temporal action proposal generation (TAPG) is a challenging task that aims to locate action instances in untrimmed videos with temporal boundaries. To evaluate the confidence of proposals, the existing works typically predict action score of proposals that are supervised by the temporal Intersection-over-Union (tIoU) between proposal and the ground-truth. In this paper, we innovatively propose a general auxiliary \emph{Background Constraint} idea to further suppress low-quality proposals, by utilizing the background prediction score to restrict the confidence of proposals. In this way, the Background Constraint concept can be easily plug-and-played into existing TAPG methods (\emph{e.g.}, BMN, GTAD). From this perspective, we propose the Background Constraint Network (BCNet) to further take advantage of the rich information of action and background. Specifically, we introduce an Action-Background Interaction module for reliable confidence evaluation, which models the inconsistency between action and background by attention mechanisms at the frame and clip levels. Extensive experiments are conducted on two popular benchmarks, \emph{i.e.}, ActivityNet-1.3 and THUMOS14. The results demonstrate that our method outperforms state-of-the-art methods. Equipped with the existing action classifier, our method also achieves remarkable performance on the temporal action localization task.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Answering complex First-Order Logical (FOL) queries on large-scale incomplete knowledge graphs (KGs) is an important yet challenging task. Recent advances embed logical queries and KG entities in the vector space and conduct query answering via dense similarity search. However, most of the designed logical operators in existing works do not satisfy the axiomatic system of classical logic. Moreover, these logical operators are parameterized so that they require a large number of complex FOL queries as training data, which are often arduous or even inaccessible to collect in most real-world KGs. In this paper, we present FuzzQE, a fuzzy logic based query embedding framework for answering FOL queries over KGs. FuzzQE follows fuzzy logic to define logical operators in a principled and learning free manner. Extensive experiments on two benchmark datasets demonstrate that FuzzQE achieves significantly better performance in answering FOL queries compared to the state-of-the-art methods. In addition, FuzzQE trained with only KG link prediction without any complex queries can achieve comparable performance with the systems trained with all FOL queries.

poster session 1 @ blue 5, poster session 11 @ blue 5, oral session 1 @ blue 5, poster session 1, poster session 11, oral session 1

Controlling neural network-based models for natural language generation (NLG) to realize desirable attributes in the generated outputs has broad applications in numerous areas such as machine translation, document summarization, and dialog systems. Approaches that enable such control in a zero-shot manner would be of great importance as, among other reasons, they remove the need for additional annotated data and training. In this work, we propose novel approaches for controlling encoder-decoder transformer-based NLG models in zero shot. While zero-shot control has previously been observed in massive models (e.g., GPT3), our method enables such control for smaller models. This is done by applying two control knobs, attention biasing and context augmentation, to these models directly during decoding and without additional training or auxiliary models. These knobs control the generation process by directly manipulating trained NLG models (e.g., biasing cross-attention layers). We show that not only are these NLG models robust to such manipulations but also their behavior could be controlled without an impact on their generation performance.

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

We study a participatory budgeting problem, where a set of strategic agents wish to split a divisible budget among different projects, by aggregating their proposals on a single division. Unfortunately, the straight-forward rule that divides the budget proportionally is susceptible to manipulation. In a recent work, (Freeman et al.,2021) proposed a class of truthful mechanisms, called moving phantom mechanisms. Among others, they propose a proportional mechanism, in the sense that in the extreme case where all agents prefer a single project to receive the whole amount, the budget is assigned proportionally. While proportionality is a naturally desired property, it is defined over a limited type of preference profiles. To address this, we expand the notion of proportionality, by proposing a quantitative framework which evaluates a budget aggregation mechanism according to its worst-case distance from the proportional allocation. Crucially, this is defined for every preference profile. We study this measure on the class of moving phantom mechanisms, and we provide approximation guarantees. For two projects, we show that the Uniform Phantom mechanism is the optimal among all truthful mechanisms. For three projects, we propose a new, proportional mechanism which is optimal among all moving phantom mechanisms. Finally, we provide impossibility results regarding the approximability of moving phantom mechanisms.

poster session 1 @ blue 6, poster session 12 @ blue 6, poster session 1, poster session 12

Weakly-supervised temporal action localization (WTAL) in untrimmed videos has emerged as a practical but challenging task since only video-level labels are available. Existing approaches typically leverage off-the-shelf segment-level features, which suffer from spatial incompleteness and temporal incoherence, thus limiting their performance. In this paper, we tackle this problem from a new perspective by enhancing segment-level representations with a simple yet effective graph convolutional network, namely action complement graph network (ACGNet). It facilitates the current video segment to perceive spatial-temporal dependencies from others that potentially convey complementary clues, implicitly mitigating the negative effects caused by the two issues above. By this means, the segment-level features are more discriminative and robust to spatial-temporal variations, contributing to higher localization accuracies. More importantly, the proposed ACGNet works as a universal module that can be flexibly plugged into different WTAL frameworks, while maintaining the end-to-end training fashion. Extensive experiments are conducted on the THUMOS'14 and ActivityNet1.2 benchmarks, where the state-of-the-art results clearly demonstrate the superiority of the proposed approach.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

We consider language modelling (LM) as a multi-label structured prediction task by re-framing training from solely predicting a single ground-truth word to ranking a set of words which could continue a given context. To avoid annotating top-k ranks, we generate them using pre-trained LMs: GPT-2, BERT, and Born-Again models. This leads to a rank-based form of knowledge distillation (KD). We also develop a method using $N$-grams to create a non-probabilistic teacher which generates the ranks without the need of a pre-trained LM. We confirm the hypotheses: that we can treat LMing as a ranking task and that we can do so without the use of a pre-trained LM. We show that rank-based KD generally gives a modest improvement to perplexity (PPL) -- though often with statistical significance -- when compared to Kullback–Leibler-based KD. Surprisingly, given the naivety of the method, the $N$-grams act as competitive teachers and achieve similar performance as using either BERT or a Born-Again model teachers. Unsurprisingly, GPT-2 always acts as the best teacher. Using it and a Transformer-XL student on Wiki-02, rank-based KD reduces a cross-entropy baseline from 65.27 to 55.94 and against a KL-based KD of 56.70.

poster session 1 @ red 4, poster session 8 @ red 4, oral session 8 @ red 4, poster session 1, poster session 8, oral session 8

We present a novel approach to hierarchical reinforcement learning for linearly-solvable Markov decision processes. Our approach assumes that the state space is partitioned, and defines subtasks for moving between the partitions. We represent value functions on several levels of abstraction, and use the compositionality of subtasks to estimate the optimal values of the states in each partition. The policy is implicitly defined on these optimal value estimates, rather than being decomposed among the subtasks. As a consequence, our approach can learn the globally optimal policy, and does not suffer from non-stationarities induced by high-level decisions. If several partitions have equivalent dynamics, the subtasks of those partitions can be shared. We show that our approach is significantly more sample efficient than that of a flat learner and similar hierarchical approaches when the set of boundary states is smaller than the entire state space.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

This paper defines fair principal component analysis (PCA) as minimizing the maximum mean discrepancy (MMD) between the dimensionality-reduced conditional distributions of different protected classes. The incorporation of MMD naturally leads to an exact and tractable mathematical formulation of fairness with good statistical properties. We formulate the problem of fair PCA subject to MMD constraints as a non-convex optimization over the Stiefel manifold and solve it using the Riemannian Exact Penalty Method with Smoothing (REPMS; Liu and Boumal, 2019). Importantly, we provide a local optimality guarantee and explicitly show the theoretical effect of each hyperparameter in practical settings, extending previous results. Experimental comparisons based on synthetic and UCI datasets show that our approach outperforms prior work in explained variance, fairness, and runtime.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

We argue that the dominant approach to explainable AI for explaining image classification, annotating images with heatmaps, provides little value for users unfamiliar with deep learning. We argue that explainable AI for images should produce output like experts produce when communicating with one another, with apprentices, and with novices. We provide an expanded set of goals of explainable AI systems and propose a Turing Test for explainable AI.

poster session 1 @ red 5, poster session 10 @ red 5, oral session 10 @ red 5, poster session 1, poster session 10, oral session 10

Accurate protein contact map prediction (PCMP) is essential for precise protein structure estimation and further biological studies. Recent works achieve significant performance on this task with high quality multiple sequence alignment (MSA). However, the PCMP accuracy drops dramatically while only poor MSA (e.g., absolute MSA count less than 10) is available. Therefore, in this paper, we propose the Contact-Distil to improve the low homologous PCMP accuracy through knowledge distillation on a self-supervised model. Particularly, two pre-trained transformers are exploited to learn the high quality and low quality MSA representation in parallel for the teacher and student model correspondingly. Besides, the co-evolution information is further extracted from pure sequence through a pretrained ESM-1b model, which provides auxiliary knowledge to improve student performance. Extensive experiments show Contact-Distil outperforms previous state-of-the-arts by large margins on CAMEO-L dataset for low homologous PCMP, i.e., around 13.3% and 9.5% improvements against Alphafold2 and MSA Transformer respectively when MSA count less than 10.

poster session 1 @ red 2, poster session 10 @ red 2, poster session 1, poster session 10

Modern generative models achieve excellent quality in a variety of tasks including image or text generation and chemical molecule modeling. However, existing methods often lack the essential ability to generate examples with requested properties, such as the age of the person in the photo or the weight of the generated molecule. Incorporating such additional conditioning factors would require rebuilding the entire architecture and optimizing the parameters from scratch. Moreover, it is difficult to disentangle selected attributes so that to perform edits of only one attribute while leaving the others unchanged. To overcome these limitations we propose PluGeN (Plugin Generative Network), a simple yet effective generative technique that can be used as a plugin to pre-trained generative models. The idea behind our approach is to transform the entangled latent representation using a flow-based module into a multi-dimensional space where the values of each attribute are modeled as an independent one-dimensional distribution. In consequence, PluGeN can generate new samples with desired attributes as well as manipulate labeled attributes of existing examples. Due to the disentangling of the latent representation, we are even able to generate samples with rare or unseen combinations of attributes in the dataset, such as a young person with gray hair, men with make-up, or women with beards. We combined PluGeN with GAN and VAE models and applied it to conditional generation and manipulation of images and chemical molecule modeling. Experiments demonstrate that PluGeN preserves the quality of backbone models while adding the ability to control the values of labeled attributes.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

Recently, deep reinforcement learning (RL) has proven its feasibility in solving combinatorial optimization problems (COPs). The learning-to-rank techniques have been studied in the field of information retrieval. While several COPs can be formulated as the prioritization of input items, as is common in the information retrieval, it has not been fully explored how the learning-to-rank techniques can be incorporated into deep RL for COPs. In this paper, we present the learning-to-rank distillation-based COP framework, where a high-performance ranking policy obtained by RL for a COP can be distilled into a non-iterative, simple model, thereby achieving a low-latency COP solver. Specifically, we employ the approximated ranking distillation to render a score-based ranking model learnable via gradient descent. Furthermore, we use the efficient sequence sampling to improve the inference performance with a limited delay. With the framework, we demonstrate that a distilled model not only achieves comparable performance to its respective, high-performance RL, but also provides several times faster inferences. We evaluate the framework with several COPs such as priority-based task scheduling and multidimensional knapsack, demonstrating the benefits of the framework in terms of inference latency and performance.

poster session 1 @ red 2, poster session 10 @ red 2, oral session 1 @ red 2, poster session 1, poster session 10, oral session 1

In this paper, we address a crucial problem in fashion e-commerce (with respect to customer experience, as well as revenue): color variants identification, i.e., identifying fashion products that match exactly in their design (or style), but only to differ in their color. We propose a generic framework, that leverages deep visual Representation Learning at its heart, to address this problem for our fashion e-commerce platform. Our framework could be trained with supervisory signals in the form of triplets, that are obtained manually. However, it is infeasible to obtain manual annotations for the entire huge collection of data usually present in fashion e-commerce platforms, such as ours, while capturing all the difficult corner cases. But, to our rescue, interestingly we observed that this crucial problem in fashion e-commerce could also be solved by simple color jitter based image augmentation, that recently became widely popular in the contrastive Self-Supervised Learning (SSL) literature, that seeks to learn visual representations without using manual labels. This naturally led to a question in our mind: Could we leverage SSL in our use-case, and still obtain comparable performance to our supervised framework? The answer is, Yes! because, color variant fashion objects are nothing but manifestations of a style, in different colors, and a model trained to be invariant to the color (with, or without supervision), should be able to recognize this! This is what the paper further demonstrates, both qualitatively, and quantitatively, while evaluating a couple of state-of-the-art SSL techniques, and also proposing a novel method.

poster session 1 @ red 6, poster session 8 @ red 6, poster session 1, poster session 8

The CO2 capture efficiency in solvent-based carbon capture systems (CCSs) critically depends on the gas-solvent interfacial area (IA), making maximization of IA a foundational challenge in CCS design. While the IA associated with a particular CCS design can be estimated via a computational fluid dynamics (CFD) simulation, using CFD to derive the IAs associated with numerous CCS designs is prohibitively costly. Fortunately, previous works such as Deep Fluids (DF) (Kim et al., 2019) show that large simulation speedups are achievable by replacing CFD simulators with neural network (NN) surrogates that faithfully mimic the CFD simulation process. This raises the possibility of a fast, accurate replacement for a CFD simulator and therefore efficient approximation of the IAs required by CCS design optimization. Thus, here, we build on the DF approach to develop surrogates that can successfully be applied to our complex carbon-capture CFD simulations. Our optimized DF-style surrogates produce large speedups (4000x) while obtaining IA relative errors as low as 4% on unseen CCS configurations that lie within the range of training configurations. This hints at the promise of NN surrogates for our CCS design optimization problem. Nonetheless, DF has inherent limitations with respect to CCS design (e.g., limited transferability of trained models to new CCS packings). We conclude with ideas to address these challenges.

poster session 1 @ red 6, poster session 9 @ red 6, poster session 1, poster session 9

In contexts where debate and deliberation is the norm, participants are regularly presented with new information that conflicts with their original beliefs. When required to update their beliefs (belief alignment), they may choose arguments that align with their worldview (confirmation bias). We test this and competing hypotheses in a constraint-based modeling approach to predict the winning arguments in multi-party interactions in the Reddit ChangeMyView dataset. We impose structural constraints that reflect competing hypotheses on a hierarchical generative Variational Auto-encoder. Our findings suggest that when arguments are further from the initial belief state of the target, they are more likely to succeed.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Geographical information provided in social media data is useful for many valuable applications. However, only a small proportion of social media posts are explicitly geotagged with their posting locations, which makes the pursuit of these applications challenging. Motivated by this, we propose a 2-level hierarchical classification method that builds upon a BERT model, coupled with textual information and temporal context, which we denote HierBERT. As far as we are aware, this work is the first to utilize a 2-level hierarchical classification approach alongside BERT and temporal information for geolocation prediction. Experimental results based on two social media datasets show that HierBERT outperforms various state-of-art baselines in terms of accuracy and distance error metrics.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

The doctoral work summarized here is an application of Artificial Intelligence (AI) for social good. The successful implementation would contribute towards low-cost, faster monitoring of road conditions across different nations, resulting in safer roads for everyone. Additionally, the study provides recommendations for re-using the road image data and the Deep Learning models released by any country for detecting road damage in other countries.

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

Person re-identification (Re-ID) based on unsupervised domain adaptation (UDA) aims to transfer the pre-trained model from one labeled source domain to an unlabeled target domain. Existing methods tackle this problem by using clustering methods to generate pseudo labels. However, pseudo labels produced by these techniques may be unstable and noisy, substantially deteriorating models' performance. In this paper, we propose a Reliability Exploration with Self-ensemble Learning (RESL) framework for domain adaptive person Re-ID. First, to increase the feature diversity, multiple branches are presented to extract features from different data augmentations. Taking the temporally average model as a mean teacher model, online label refining is conducted by using its dynamic ensemble predictions from different branches as soft labels. Second, to combat the adverse effects of unreliable samples in clusters, sample reliability is estimated by evaluating the consistency of different clusters’ results, followed by selecting reliable instances for training and re-weighting sample contribution within Re-ID losses. A contrastive loss is also utilized with cluster-level memory features which are updated by the mean feature. The experiments demonstrate that our method can significantly surpass the state-of-the-art performance on the unsupervised domain adaptive person Re-ID.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

The shortcomings of maximum likelihood estimation in the context of model-based reinforcement learning have been highlighted by an increasing number of papers. When the model class is misspecified or has a limited representational capacity, model parameters with high likelihood might not necessarily result in high performance of the agent on a downstream control task. To alleviate this problem, we propose an end-to-end approach for model learning which directly optimizes the expected returns using implicit differentiation. We treat a value function that satisfies the Bellman optimality operator induced by the model as an implicit function of model parameters and show how to differentiate the function. We provide theoretical and empirical evidence highlighting the benefits of our approach in the model misspecification regime compared to likelihood-based methods.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

We study the computational complexity of computing solutions for the straight-cut and square-cut pizza sharing problems. We show that finding an approximate solution is PPA-hard for the straight-cut problem, and PPA-complete for the square-cut problem, while finding an exact solution for the square-cut problem is FIXP-hard and in BU. Our PPA-hardness results apply even when all mass distributions are unions of non-overlapping squares, and our FIXP-hardness result applies even when all mass distributions are unions of weighted squares and right-angled triangles. We also prove that decision variants of the square-cut problem are hard: the approximate problem is NP-complete, and the exact problem is ETR-complete.

poster session 1 @ blue 6, poster session 12 @ blue 6, poster session 1, poster session 12

In this work, we introduce 6D Convolutional Neural Network (6DCNN) designed to tackle the problem of detecting relative positions and orientations of local patterns when processing three-dimensional volumetric data. 6DCNN also includes SE(3)-equivariant message-passing and nonlinear activation operations constructed in the Fourier space. Working in the Fourier space allows significantly reducing the computational complexity of our operations. We demonstrate the properties of the 6D convolution and its efficiency in the recognition of spatial patterns. We also assess the 6DCNN model on several datasets from the recent CASP protein structure prediction challenges. Here, 6DCNN improves over the baseline architecture and also outperforms the state of the art.

poster session 1 @ red 2, poster session 10 @ red 2, poster session 1, poster session 10

Temporal action detection aims to locate the boundaries of action in the video. The current method based on boundary matching enumerates and calculates all possible boundary matchings to generate proposals. However, these methods neglect the long-range context aggregation in boundary prediction. At the same time, due to the similar semantics of adjacent matchings, local semantic aggregation of densely-generated matchings cannot improve semantic richness and discrimination. In this paper, we propose the end-to-end proposal generation method named Dual Context Aggregation Network (DCAN) to aggregate context on two levels, namely, boundary level and proposal level, for generating high-quality action proposals, thereby improving the performance of temporal action detection. Specifically, we design the Multi-Path Temporal Context Aggregation (MTCA) to achieve smooth context aggregation on boundary level and precise evaluation of boundaries. For matching evaluation, Coarse-to-fine Matching (CFM) is designed to aggregate context on the proposal level and refine the matching map from coarse to fine. We conduct extensive experiments on ActivityNet v1.3 and THUMOS-14. DCAN obtains an average mAP of 35.39% on ActivityNet v1.3 and reaches mAP 54.14% at IoU@0.5 on THUMOS-14, which demonstrates DCAN can generate high-quality proposals and achieve state-of-the-art performance. We release the code at https://github.com/cg1177/DCAN.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Domain generalization typically requires data from multiple source domains for model learning. However, such strong assumption may not always hold in practice, especially in medical field where the data sharing is highly concerned and sometimes prohibitive due to privacy issue. This paper studies the important yet challenging single domain generalization problem, in which a model is learned under the worst-case scenario with only one source domain to directly generalize to different unseen target domains. We present a novel approach to address this problem in medical image segmentation, which extracts and integrates the semantic shape prior information of segmentation that are invariant across domains and can be well-captured even from single domain data to facilitate segmentation under distribution shifts. Besides, a test-time adaptation strategy with dual-consistency regularization is further devised to promote dynamic incorporation of these shape priors under each unseen domain to improve model generalizability. Extensive experiments on two medical image segmentation tasks demonstrate the consistent improvements of our method across various unseen domains, as well as its superiority over state-of-the-art approaches in addressing domain generalization under the worst-case scenario.

poster session 1 @ red 2, poster session 10 @ red 2, poster session 1, poster session 10

Integer programs provide a powerful abstraction for representing a wide range of real-world scheduling problems. Despite their ability to model general scheduling problems, solving large-scale integer programs (IP) remains a computational challenge in practice. The incorporation of more complex objectives such as robustness to disruptions further exacerbates the computational challenge. We present NICE (Neural network IP Coefficient Extraction), a novel technique that combines reinforcement learning and integer programming to tackle the problem of robust scheduling. More specifically, NICE uses reinforcement learning to approximately represent complex objectives in an integer programming formulation. We use NICE to determine assignments of pilots to a flight crew schedule so as to reduce the impact of disruptions. We compare NICE with (1) a baseline integer programming formulation that produces a feasible crew schedule, and (2) a robust integer programming formulation that explicitly tries to minimize the impact of disruptions. Our experiments show that, across a variety of scenarios, NICE produces schedules resulting in 33% to 48% fewer disruptions than the baseline formulation. Moreover, in more severely constrained scheduling scenarios in which the robust integer program fails to produce a schedule within 90 minutes, NICE is able to build robust schedules in less than 2 seconds on average.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Crime prediction plays an impactful role in enhancing public security and sustainable development of urban. With recent advances in data collection and integration technologies, a large amount of urban data with rich crime-related information and fine-grained spatio-temporal logs have been recorded. Such helpful information can boost our understandings of the temporal evolution and spatial factors of urban crimes and can enhance accurate crime prediction. However, the vast majority of existing crime prediction algorithms either do not distinguish different types of crime or treat each crime type separately, which fails to capture the intrinsic correlations among different types of crime. In this paper, we perform crime prediction exploiting the cross-type and spatio-temporal correlations of urban crimes. In particular, we verify the existence of correlations among different types of crime from temporal and spatial perspectives, and propose a coherent framework to mathematically model these correlations for crime prediction. Extensive experiments on real-world datasets validate the effectiveness of our framework.

poster session 1 @ red 2, poster session 10 @ red 2, poster session 1, poster session 10

More than US$ 27 billion is estimated to have been paid-out in farm support in USA alone since 1991 in response to climate change impacts on agriculture, with costs likely continuing to rise. With the wider adoption of precision agriculture - an agriculture management strategy that involves gathering, processing and analyzing temporal, spatial and individual data - in both developed and developing countries, there is an increasing opportunity to harness accumulating, shareable, big data using artificial intelligence (AI) methods, collected from weather stations, field sensor networks, Internet-of-Things devices, unmanned aerial vehicles, and earth observational satellites. This requires smart algorithms tailored to agricultural data types, integrated into digital solutions that are viable, flexible, and scalable for wide deployment for a wide variety of agricultural users and decision-makers. We discuss a novel AI approach that addresses the real-world problem of developing a viable solution for reliably, timely, and cost-effectively forecasting crop status across large agricultural regions using Earth observational information in near-real-time. Our approach is based on extracting time-conditioned topological features which characterize complex spatio-temporal dependencies between crop production regions and integrating such topological signatures into Long Short Term Memory (LSTM). We discuss utility and limitations of the resulting zigzag persistence-based LSTM (ZZTop-LSTM) as a new tool for developing more informed crop insurance rate-making and accurate tracking of changing risk exposures and vulnerabilities within insurance risk areas.

poster session 1 @ red 6, poster session 9 @ red 6, poster session 1, poster session 9

Due to difficulties in acquiring ground truth depth of equirectangular (360) images, the quality and quantity of equirectangular depth data today is insufficient to represent the various scenes in the world. Therefore, 360 depth estimation studies, which relied solely on supervised learning, are destined to produce unsatisfactory results. Although self-supervised learning methods focusing on equirectangular images (EIs) are introduced, they often have incorrect or non-unique solutions, causing unstable performance. In this paper, we propose 360 monocular depth estimation methods which improve on the areas that limited previous studies. First, we introduce a self-supervised 360 depth learning method that only utilizes gravity-aligned videos, which has the potential to eliminate the needs for depth data during the training procedure. Second, we propose a joint learning scheme realized by combining supervised and self-supervised learning. The weakness of each learning is compensated, thus leading to more accurate depth estimation. Third, we propose a non-local fusion block, which can further retain the global information encoded by vision transformer when reconstructing the depths. With the proposed methods, we successfully apply the transformer to 360 depth estimations, to the best of our knowledge, which has not been tried before. On several benchmarks, our approach achieves significant improvements over previous works and establishes a state of the art.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Video anomaly detection aims to automatically identify unusual objects or behaviours by learning from normal videos. Previous methods tend to use simplistic reconstruction or prediction constraints, which leads to the insufficiency of learned representations for normal data. As such, we propose a novel bi-directional architecture with three consistency constraints to comprehensively regularize the prediction task from pixel-wise, cross-modal, and temporal-sequence levels. First, predictive consistency is proposed to consider the symmetry property of motion and appearance in forwards and backwards time, which ensures the highly realistic appearance and motion predictions at the pixel-wise level. Second, association consistency considers the relevance between different modalities and uses one modality to regularize the prediction of another one. Finally, temporal consistency utilizes the relationship of the video sequence and ensures that the predictive network generates temporally consistent frames. During inference, the pattern of abnormal frames is unpredictable and will therefore cause higher prediction errors. Experiments show that our method outperforms advanced anomaly detectors and achieves state-of-the-art results on UCSD Ped2, CUHK Avenue, and ShanghaiTech datasets.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Learning visual knowledge from massive weakly-labeled web videos has attracted growing research interests thanks to the large corpus of easily accessible video data on the Internet. However, for video action recognition, the action of interest might only exist in arbitrary clips of untrimmed web videos, resulting in high label noises in the temporal space. To address this challenge, we introduce a new method for pre-training video action recognition models using queried web videos. Instead of trying to filter out potential noises, we propose to provide fine-grained supervision signals by defining the concept of Sub-Pseudo Label (SPL). Specifically, SPL spans out a new set of meaningful "middle ground" label space constructed by extrapolating the original weak labels during video querying and the prior knowledge distilled from a teacher model. Consequently, SPL provides enriched supervision for video models to learn better representations and improves data utilization efficiency of untrimmed videos. We validate the effectiveness of our method on four video action recognition datasets and a weakly-labeled image dataset. Experiments show that SPL outperforms several existing pre-training strategies and the learned representations lead to competitive results on several benchmarks.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Intensive care in hospitals is distributed to different units that care for patient populations reflecting specific comorbidities, treatments, and outcomes. Unit expertise can be shared to potentially improve the quality of methods and outcomes for patients across units. We propose an algorithmic rule pruning approach for use in building short lists of human-interpretable rules that reliably identify patient beneficiaries of expertise transfers in the form of machine learning risk models. Our experimental results, obtained with two intensive care monitoring datasets, demonstrate the potential utility of the proposed method in practice.

poster session 1 @ red 6, poster session 2 @ red 6, poster session 1, poster session 2

A common approach to Temporal Action Localization (TAL) is to generate action proposals and then perform action classification and localization on them. For each proposal, existing methods universally use a shared proposal-level representation for both tasks. However, our analysis indicates that this shared representation focuses on the most discriminative frames for classification, e.g., ``take-offs" rather than ``run-ups" in distinguishing ``high jump" and ``long jump", while frames most relevant to localization, such as the start and end frames of an action, are largely ignored. In other words, such a shared representation can not simultaneously handle both classification and localization tasks well, and it makes precise TAL difficult. To address this challenge, this paper disentangles the shared representation into classification and localization representations. The disentangled classification representation focuses on the most discriminative frames, and the disentangled localization representation focuses on the action phase as well as the action start and end. Our model could be divided into two sub-networks, i.e., the disentanglement network and the context-based aggregation network. The disentanglement network is an autoencoder to learn orthogonal hidden variables of classification and localization. The context-based aggregation network aggregates the classification and localization representations by modeling local and global contexts. We evaluate our proposed method on two popular benchmarks for TAL, which outperforms all state-of-the-art methods.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

In this paper, we give a new definition for sample complexity, and further develop a theoretical analysis to bridge the gap between sample complexity and model capacity. In contrast to previous works which study on some toy samples, we conduct our analysis on more general data space, and build a qualitative relationship from sample complexity to model capacity required to achieve comparable performance. Besides, we introduce a simple indicator to evaluate the sample complexity based on continuous mapping. Moreover, we further analysis the relationship between sample complexity and data distribution, which paves the way to understand the present representation learning. Extensive experiments on several datasets well demonstrate the effectiveness of our evaluation method.

poster session 1 @ blue 4, poster session 11 @ blue 4, oral session 11 @ blue 4, poster session 1, poster session 11, oral session 11

Classical measurements and modelling that underpin present flood warning and alert systems are based on fixed and spatially restricted static sensor networks. Computationally expensive physics-based simulations are often used that can't react in real-time to changes in environmental conditions. We want to explore contemporary artificial intelligence (AI) for predicting flood risk in real time by using a diverse range of data sources. By combining heterogeneous data sources, we aim to nowcast rapidly changing flood conditions and gain a grater understanding of urgent humanitarian needs.

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

Transformers achieve state-of-the-art performance for natural language processing tasks by pre-training on large-scale text corpora. They are extremely compute-intensive and have very high sample complexity. Memory replay is a mechanism that remembers and reuses past examples by saving to and replaying from a memory buffer. It has been successfully used in reinforcement learning and GANs due to better sample efficiency. In this paper, we propose Transformer with Memory Replay, which integrates memory replay with transformer, making transformer more sample efficient. Experiments on GLUE and SQuAD benchmark datasets showed that Transformer with Memory Replay can achieve at least 1% point increase compared to the baseline transformer model when pre-trained with the same number of examples. Further, by adopting a careful design that reduces the wall-clock time overhead of memory replay, we also empirically achieve a better runtime efficiency.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Product copywriting is a critical component of e-commerce recommendation platforms. It aims to attract users' interest and improve user experience by highlighting product characteristics with textual descriptions. In this paper, we report our experience deploying the proposed Automatic Product Copywriting Generation (APCG) system into the JD.com e-commerce product recommendation platform. It consists of two main components: 1) natural language generation, which is built from a transformer-pointer network and a pre-trained sequence-to-sequence model based on millions of training data from our in-house platform; and 2) copywriting quality control, which is based on both automatic evaluation and human screening. For selected domains, the models are trained and updated daily with the updated training data. In addition, the model is also used as a real-time writing assistant tool on our live broadcast platform. The APCG system has been deployed in JD.com since Feb 2021. By Sep 2021, it has generated 2.53 million product descriptions, and improved the overall averaged click-through rate (CTR) and the Conversion Rate (CVR) by 4.22% and 3.61%, compared to baselines, respectively on a year-on-year basis. The accumulated Gross Merchandise Volume (GMV) made by our system is improved by 213.42%, compared to the number in Feb 2021.

poster session 1 @ red 6, poster session 8 @ red 6, poster session 1, poster session 8

Knowledge Distillation (KD) is a widely-used technology to inherit information from cumbersome teacher models to compact student models, consequently realizing model compression and acceleration. Compared with image classification, object detection is a more complex task, and designing specific KD methods for object detection is non-trivial. In this work, we elaborately study the behaviour difference between the teacher and student detection models, and obtain two intriguing observations: First, the teacher and student rank their detected candidate boxes quite differently, which results in their precision discrepancy. Second, there is a considerable gap between the feature response differences and prediction differences between teacher and student, indicating that equally imitating all the feature maps of the teacher is the sub-optimal choice for improving the student's accuracy. Based on the two observations, we propose Rank Mimicking (RM) and Prediction-guided Feature Imitation (PFI) for distilling one-stage detectors, respectively. RM takes the rank of candidate boxes from teachers as a new form of knowledge to distill, which consistently outperforms the traditional soft label distillation. PFI attempts to correlate feature differences with prediction differences, making feature imitation directly help to improve the student's accuracy. On MS COCO and PASCAL VOC benchmarks, extensive experiments are conducted on various detectors with different backbones to validate the effectiveness of our method. Specifically, RetinaNet with ResNet50 achieves 40.4% mAP in MS COCO, which is 3.5% higher than its baseline, and also outperforms previous KD methods.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Recent studies in the hate and counter hate domain have provided the grounds for investigating how to detect this pervasive content in social media. These studies mostly work with synthetic replies to hateful content written by annotators on demand rather than replies written by real users. We argue that working with naturally occurring replies to hateful content is key to study the problem. Building on this motivation, we create a corpus of 5,652 hateful tweets and replies. We analyze their fine-grained relationships by indicating whether the reply (a) is hate or counter hate speech, (b) provides a justification, (c) attacks the author of the tweet, and (d) adds additional hate. We also present linguistic insights into the language people use depending on these fine-grained relationships. Experimental results show improvements (a) taking into account the hateful tweet in addition to the reply and (b) pretraining with related tasks.

poster session 1 @ red 4, poster session 8 @ red 4, oral session 8 @ red 4, poster session 1, poster session 8, oral session 8

Attention mechanism has been widely believed as the key to success of vision transformers (ViTs), since it provides a flexible and powerful way to model spatial relationships. However, is the attention mechanism truly an indispensable part of ViT? Can it be replaced by some other alternatives? To demystify the role of attention mechanism, we simplify it into an extremely simple case: ZERO FLOP and ZERO parameter. Concretely, we revisit the shift operation. It does not contain any parameter or arithmetic calculation. The only operation is to exchange a small portion of the channels between neighboring features. Based on this simple operation, we construct a new backbone network, namely ShiftViT, where the attention layers in ViT are substituted by shift operations. Surprisingly, ShiftViT works quite well in several mainstream tasks, e.g., classification, detection, and segmentation. The performance is on par with or even better than the strong baseline Swin Transformer. These results suggest that the attention mechanism might not be the vital factor that makes ViT successful. It can be even replaced by a zero-parameter operation. We should pay more attentions to the remaining parts of ViT in the future work. Code is available at github.com/microsoft/SPACH.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Learning effective representations of entities and relations for knowledge graphs (KGs) is critical to the success of many multi-relational learning tasks. Existing methods based on graph neural networks learn a deterministic embedding function, which lacks sufficient flexibility to explore better choices when dealing with the imperfect and noisy KGs such as the scarce labeled nodes and noisy graph structure. To this end, we propose a novel multi-relational graph Gaussian Process network (GGPN), which aims to improve the flexibility of deterministic methods by simultaneously learning a family of embedding functions, i.e., a stochastic embedding function. Specifically, a Bayesian Gaussian Process (GP) is proposed to model the distribution of this stochastic function and the resulting representations are obtained by aggregating stochastic function values, i.e., messages, from neighboring entities. The two problems incurred when leveraging GP in GGPN are the proper choice of kernel function and the cubic computational complexity. To address the first problem, we further propose a novel kernel function that can explicitly take the diverse relations between each pair of entities into account and be adaptively learned in a data-driven way. We address the second problem by reformulating GP as a Bayesian linear model, resulting in a linear computational complexity. With these two solutions, our GGPN can be efficiently trained in an end-to-end manner. We evaluate our GGPN in link prediction and entity classification tasks, and the experimental results demonstrate the superiority of our method.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Cross-modal person re-identification (Re-ID) is critical for modern video surveillance systems. The key challenge is to align cross-modality representations conforming to semantic information present for a person and ignore background information. This work presents a novel convolutional neural network (CNN) based architecture designed to learn semantically aligned cross-modal visual and textual representations. The underlying building block, named AXM-Block, is a unified multi-layer network that dynamically exploits the multi-scale knowledge from both modalities and re-calibrates each modality according to shared semantics. To complement the convolutional design, contextual attention is applied in the text branch to manipulate long-term dependencies. Moreover, we propose a unique design to enhance visual part-based feature coherence and locality information. Our framework is novel in its ability to implicitly learn aligned semantics between modalities during the feature learning stage. The unified feature learning effectively utilizes textual data as a super-annotation signal for visual representation learning and automatically rejects irrelevant information. The entire AXM-Net is trained end-to-end on CUHK-PEDES data. We report results on two tasks, person search and cross-modal Re-ID. The AXM-Net outperforms the current state-of-the-art (SOTA) methods and achieves 64% Rank@1 on the CUHK-PEDES test set. It also outperforms by >10% for cross-viewpoint text-to-image Re-ID scenarios on CrossRe-ID and CUHK-SYSU datasets.

poster session 1 @ red 2, poster session 10 @ red 2, oral session 10 @ red 2, poster session 1, poster session 10, oral session 10

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

We present Neural Marionette, an unsupervised approach that discovers the skeletal structure from a dynamic sequence and learns to generate diverse motions that are consistent with the observed motion dynamics. Given a video stream of point cloud observation of an articulated body in arbitrary poses, our approach discovers the unknown low-dimensional skeletal relationship that can effectively represent the motion. Then the discovered structure is utilized to encode the motion priors of dynamic sequences in a latent structure, which can be decoded to the relative joint rotations to represent the full skeletal motion. Our approach works without any prior knowledge of the underlying motion or skeletal structure, and we demonstrate that the discovered structure is even comparable to the hand-labeled ground truth skeleton in representing a 4D sequence of motion. The skeletal structure embeds the general semantics of possible motion space that can generate motions for diverse scenarios. We verify that the learned motion prior is generalizable to the multi-modal sequence generation, interpolation of two poses, and motion retargeting to a different skeletal structure.

poster session 1 @ red 1, poster session 11 @ red 1, oral session 1 @ red 1, poster session 1, poster session 11, oral session 1

Robust adversarial reinforcement learning is an effective method to train agents to manage uncertain disturbance and modeling errors in real environments. However, for systems that are sensitive to disturbances or those that are difficult to stabilize, it is easier to learn a powerful adversary than establish a stable control policy. An improper strong adversary can destabilize the system, introduce biases in the sampling process, make the learning process unstable, and even reduce the robustness of the policy. In this study, we consider the problem of ensuring system stability during training in the adversarial reinforcement learning architecture. The dissipative principle of robust $H_\infty$ control is extended to the Markov Decision Process, and robust stability constraints are obtained based on $L_2$ gain performance in the reinforcement learning system. Thus, we propose a dissipation-inequation-constraint-based adversarial reinforcement learning architecture. This architecture ensures the stability of the system during training by imposing constraints on the normal and adversarial agents. Theoretically, this architecture can be applied to a large family of deep reinforcement learning algorithms. Results of experiments in MuJoCo and GymFc environments show that our architecture effectively improves the robustness of the controller against environmental changes and adapts to more powerful adversaries. Results of the flight experiments on a real quadcopter indicate that our method can directly deploy the policy trained in the simulation environment to the real environment, and our controller outperforms the PID controller based on hardware-in-the-loop. Both our theoretical and empirical results provide new and critical outlooks on the adversarial reinforcement learning architecture from a rigorous robust control perspective.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

Online reinforcement learning (RL) algorithms are often difficult to deploy in complex human-facing applications as they may learn slowly and have poor early performance. To address this, we introduce a practical algorithm for incorporating human insight to speed learning. Our algorithm, Constraint Sampling Reinforcement Learning (CSRL), incorporates prior domain knowledge as constraints/restrictions on the RL policy. It takes in multiple potential policy constraints to maintain robustness to misspecification of individual constraints while leveraging helpful ones to learn quickly. Given a base RL learning algorithm (ex. UCRL, DQN, Rainbow) we propose an upper confidence with elimination scheme that leverages the relationship between the constraints, and their observed performance, to adaptively switch among them. We instantiate our algorithm with DQN-type algorithms and UCRL as base algorithms, and evaluate our algorithm in four environments, including three simulators based on real data: recommendations, educational activity sequencing, and HIV treatment sequencing. In all cases, CSRL learns a good policy faster than baselines.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

Automatic delineation of organ-at-risk (OAR) and gross-tumor-volume (GTV) is of great significance for radiotherapy planning. However, it is a challenging task to learn powerful representations for accurate delineation under limited pixel (voxel)-wise annotations. Contrastive learning at pixel-level can alleviate the dependency on annotations by learning dense representations from unlabeled data. Recent studies in this direction design various contrastive losses on the feature maps, to yield discriminative features for each pixel in the map. However, pixels in the same map inevitably share semantics to be closer than they actually are, which may affect the discrimination of pixels in the same map and lead to the unfair comparison to pixels in other maps. To address these issues, we propose a separated region-level contrastive learning scheme, namely SepaReg, the core of which is to separate each image into regions and encode each region separately. Specifically, SepaReg comprises two components: a structure-aware image separation (SIS) module and an intra- and inter-organ distillation (IID) module. The SIS is proposed to operate on the image set to rebuild a region set under the guidance of structural information. The inter-organ representation will be learned from this set via typical contrastive losses cross regions. On the other hand, the IID is proposed to tackle the quantity imbalance in the region set as tiny organs may produce fewer regions, by exploiting intra-organ representations. We conducted extensive experiments to evaluate the proposed model on a public dataset and two private datasets. The experimental results demonstrate the effectiveness of the proposed model, consistently achieving better performance than state-of-the-art approaches. Code is available at https://github.com/jcwang123/Separate_CL.

poster session 1 @ red 2, poster session 10 @ red 2, oral session 10 @ red 2, poster session 1, poster session 10, oral session 10

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

Heads-up no-limit Texas hold'em (HUNL) is the quintessential game with imperfect information. Representative prior works like DeepStack and Libratus heavily rely on counterfactual regret minimization (CFR) and its variants to tackle HUNL. However, the prohibitive computation cost of CFR iteration makes it difficult for subsequent researchers to learn the CFR model in HUNL and apply it in other practical applications. In this work, we present AlphaHoldem, a high-performance and lightweight HUNL AI obtained with an end-to-end self-play reinforcement learning framework. The proposed framework adopts a pseudo-Siamese architecture to directly learn from the input state information to the output actions by competing the learned model with its different historical versions. The main technical contributions include a novel state representation of card and betting information, a multi-task self-play training loss function, and a new model evaluation and selection metric to generate the final model. In a study involving 100,000 hands of poker, AlphaHoldem defeats Slumbot and DeepStack using only one PC with three days training. At the same time, AlphaHoldem only takes four milliseconds for each decision-making using only a single CPU core, more than 1,000 times faster than DeepStack. We will provide an online testing platform of AlphaHoldem to facilitate further studies in this direction.

poster session 1 @ red 2, poster session 10 @ red 2, oral session 1 @ red 2, poster session 1, poster session 10, oral session 1

The conventional wisdom behind learning deep classification models is to focus on bad-classified examples and ignore well-classified examples that are far from the decision boundary. For instance, when training with cross-entropy loss, examples with higher likelihoods (i.e., well-classified examples) contribute smaller gradients in back-propagation. However, we theoretically show that this common practice hinders representation learning, energy optimization, and margin growth. To counteract this deficiency, we propose to reward well-classified examples with additive bonuses to revive their contribution to the learning process. This counterexample theoretically addresses these three issues. We empirically support this claim by directly verifying the theoretical results or significant performance improvement with our counterexample on diverse tasks, including image classification, graph classification, and machine translation. Furthermore, this paper shows that we can deal with complex scenarios, such as imbalanced classification, OOD detection, and applications under adversarial attacks, because our idea can solve these three issues.Code is available at \url{https://github.com/lancopku/well-classified-examples-are-underestimated}.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

We propose a new approach, the calibrated nonparametric scan statistic (CNSS), for more accurate detection of anomalous patterns in large-scale, real-world graphs. Scan statistics identify connected subgraphs that are interesting or unexpected through maximization of a likelihood ratio statistic; in particular, nonparametric scan statistics (NPSSs) identify subgraphs with a higher than expected proportion of individually significant nodes. However, we show that recently proposed NPSS methods are miscalibrated, failing to account for the maximization of the statistic over the multiplicity of subgraphs. This results in both reduced detection power for subtle signals, and low precision of the detected subgraph even for stronger signals. Thus we develop a new statistical approach to recalibrate NPSSs, correctly adjusting for multiple hypothesis testing and taking the underlying graph structure into account. While the recalibration, based on randomization testing, is computationally expensive, we propose both an efficient (approximate) algorithm and new, closed-form lower bounds (on the expected maximum proportion of significant nodes for subgraphs of a given size, under the null hypothesis of no anomalous patterns). These advances, along with the integration of recent core-tree decomposition methods, enable CNSS to scale to large real-world graphs, with substantial improvement in the accuracy of detected subgraphs. Extensive experiments on both semi-synthetic and real-world datasets are demonstrated to validate the effectiveness of our proposed methods, in comparison with state-of-the-art counterparts.

poster session 1 @ blue 5, poster session 11 @ blue 5, oral session 1 @ blue 5, poster session 1, poster session 11, oral session 1

Estimating value functions is a core component of reinforcement learning algorithms. Temporal difference (TD) learning algorithms use bootstrapping, i.e. they update the value function toward a learning target using value estimates at subsequent time-steps. Alternatively, the value function can be updated toward a learning target constructed by separately predicting successor features (SF)--a policy-dependent model--and linearly combining them with instantaneous rewards. We focus on bootstrapping targets used when estimating value functions, and propose a new backup target, the $\eta$-return mixture, which implicitly combines value-predictive knowledge (used by TD methods) with (successor) feature-predictive knowledge--with a parameter $\eta$ capturing how much to rely on each. We illustrate that incorporating predictive knowledge through an $\eta\gamma$-discounted SF model makes more efficient use of sampled experience, compared to either extreme, i.e. bootstrapping entirely on the value function estimate, or bootstrapping on the product of separately estimated successor features and instantaneous reward models. We empirically show this approach leads to faster policy evaluation and better control performance, for tabular and nonlinear function approximations, indicating scalability and generality.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

Graph comparison deals with identifying similarities and dissimilarities between graphs. A major obstacle is the unknown alignment of graphs, as well as the lack of accurate and inexpensive comparison metrics. In this work we introduce the filter graph distance. It is an optimal transport based distance which drives graph comparison through the probability distribution of filtered graph signals. This creates a highly flexible distance, capable of prioritising different spectral information in observed graphs, offering a wide range of choices for a comparison metric. We tackle the problem of graph alignment by computing graph permutations that minimise our new filter distances, which implicitly solves the graph comparison problem. We then propose a new approximate cost function that circumvents many computational difficulties inherent to graph comparison and permits the exploitation of fast algorithms such as mirror gradient descent, without grossly sacrificing the performance. We finally propose a novel algorithm derived from a stochastic version of mirror gradient descent, which accommodates the non-convexity of the alignment problem, offering a good trade-off between performance accuracy and speed. The experiments on graph alignment and classification show that the flexibility gained through filter graph distances can have a significant impact on performance, while the difference in speed offered by the approximation cost makes the framework applicable in practical settings.

poster session 1 @ blue 3, poster session 12 @ blue 3, oral session 1 @ blue 3, poster session 1, poster session 12, oral session 1

Spatio-temporal forecasting is challenging attributing to the high nonlinearity in temporal dynamics as well as complex location-characterized patterns in spatial domains, especially in fields like weather forecasting. Graph convolutions are usually used for modeling the spatial dependency in meteorology to handle the irregular distribution of sensors' spatial location. In this work, a novel graph-based convolution for imitating the meteorological flows is proposed to capture the local spatial patterns. Based on the assumption of smoothness of location-characterized patterns, we propose conditional local convolution whose shared kernel on nodes' local space is approximated by neural networks, with local representations of coordinate obtained by horizon maps into cylindrical-tangent space as its input. We further propose the distance and orientation scaling terms to reduce the impacts of irregular spatial distribution. The convolution is embedded in a Recurrent Neural Network architecture to model the temporal dynamics, leading to the Conditional Local Convolution Recurrent Network (CLCRN). Our model is evaluated on real-world weather benchmark datasets, achieving state-of-the-art performance with obvious improvements. We conduct further analysis on local pattern visualization, model's framework choice, advantages of horizon maps and etc.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

RST-style discourse parsing plays a vital role in many NLP tasks, revealing the underlying semantic/pragmatic structure of potentially complex and diverse documents. Despite its importance, one of the most prevailing limitations in modern day discourse parsing is the lack of large-scale datasets. To overcome the data sparsity issue, distantly supervised approaches from tasks like sentiment analysis and summarization have been recently proposed. Here, we extend this line of research by exploiting distant supervision from topic segmentation, which can arguably provide a strong and oftentimes complementary signal for high-level discourse structures. Experiments on two human-annotated discourse treebanks confirm that our proposal generates accurate tree structures on sentence and paragraph level, consistently outperforming previous distantly supervised models on the sentence-to-document task and occasionally reaching even higher scores on the sentence-to-paragraph level.

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

Due to the lack of expertise for medical image annotation, the investigation of label-efficient methodology for medical image segmentation becomes a heated topic. Recent progresses focus on the efficient utilization of weak annotations together with few strongly-annotated labels so as to achieve comparable segmentation performance in many unprofessional scenarios. However, these approaches only concentrate on the supervision inconsistency between strongly- and weakly-annotated instances but ignore the instance inconsistency inside the weakly-annotated instances, which inevitably leads to performance degradation. To address this problem, we propose a novel label-efficient hybrid-supervised framework, which considers each weakly-annotated instance individually and learns its weight guided by the gradient direction of the strongly-annotated instances, so that the high-quality prior in the strongly-annotated instances is better exploited and the weakly-annotated instances are depicted more precisely. Specially, our designed dynamic instance indicator (DII) realizes the above objectives, and is adapted to our dynamic co-regularization (DCR) framework further to alleviate the erroneous accumulation from distortions of weak annotations. Extensive experiments on two hybrid-supervised medical segmentation datasets demonstrate that with only 10% strong labels, the proposed framework can leverage the weak labels efficiently and achieve competitive performance against the 100% strong-label supervised scenario.

poster session 1 @ red 2, poster session 10 @ red 2, poster session 1, poster session 10

We present PantheonRL, a multiagent reinforcement learning software package for dynamic training interactions such as round-robin, adaptive, and ad-hoc training. Our package is designed around flexible agent objects that can be easily configured to support different training interactions, and handles fully general multiagent environments with mixed rewards and n agents. Built on top of StableBaselines3, our package works directly with existing powerful deep RL algorithms. Finally, PantheonRL comes with an intuitive yet functional web user interface for configuring experiments and launching multiple asynchronous jobs. Our package can be found at https://github.com/Stanford-ILIAD/PantheonRL.

poster session 1 @ red 5, poster session 10 @ red 5, poster session 1, poster session 10

We present a novel dynamic recommendation model that focuses on users who have interactions in the past but turn relatively inactive recently. Making effective recommendations to these {\em time-sensitive cold-start users} is critical to maintain the user base of a recommender system. Due to the sparse recent interactions, it is challenging to capture these users' current preferences precisely. Solely relying on their historical interactions may also lead to outdated recommendations misaligned with their recent interests. The proposed model leverages historical and current user-item interactions and dynamically factorizes a user's (latent) preference into time-specific and time-evolving representations that jointly affect user behaviors. These latent factors further interact with an optimized item embedding to achieve accurate and timely recommendations. Experiments over real-world data help demonstrate the effectiveness of the proposed time-sensitive cold-start recommendation model.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

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

Few-shot object detection (FSOD) aims to detect objects using only a few examples. How to adapt state-of-the-art object detectors to the few-shot domain remains challenging. Object proposal is a key ingredient in modern object detectors. However, the quality of proposals generated for few-shot classes using existing methods is far worse than that of many-shot classes, e.g., missing boxes for few-shot classes due to misclassification or inaccurate spatial locations with respect to true objects. To address the noisy proposal problem, we propose a novel meta-learning based FSOD model by jointly optimizing the few-shot proposal generation and fine-grained few-shot proposal classification. To improve proposal generation for few-shot classes, we propose to learn a lightweight metric-learning based prototype matching network, instead of the conventional simple linear object/nonobject classifier, e.g., used in RPN. Our non-linear classifier with the feature fusion network could improve the discriminative prototype matching and the proposal recall for few-shot classes. To improve the fine-grained few-shot proposal classification, we propose a novel attentive feature alignment method to address the spatial misalignment between the noisy proposals and few-shot classes, thus improving the performance of few-shot object detection. Meanwhile we learn a separate Faster R-CNN detection head for many-shot base classes and show strong performance of maintaining base-classes knowledge. Our model achieves state-of-the-art performance on multiple FSOD benchmarks over most of the shots and metrics.

poster session 1 @ red 3, poster session 11 @ red 3, oral session 11 @ red 3, poster session 1, poster session 11, oral session 11

The task of Legal Statute Identification (LSI) aims to identify the legal statutes that are relevant to a given description of facts or evidence of a legal case. Existing methods only utilize the textual content of facts and legal articles to guide such a task. However, the citation network among case documents and legal statutes is a rich source of additional information, which is not considered by existing models. In this work, we take the first step towards utilising both the text and the legal citation network for the LSI task. We curate a large novel dataset for this task, including facts of cases from several major Indian Courts of Law, and statutes from the Indian Penal Code (IPC). Modeling the statutes and training documents as a heterogeneous graph, our proposed model LeSICiN can learn rich textual and graphical features, and can also tune itself to correlate these features. Thereafter, the model can be used to inductively predict links between test documents (new nodes whose graphical features are not available to the model) and statutes (existing nodes). Extensive experiments on the dataset show that our model comfortably outperforms several state-of-the-art baselines, by exploiting the graphical structure along with textual features.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

We study the unsupervised representation learning for the semantic segmentation task. Different from previous works that aim at providing unsupervised pre-trained backbones for segmentation models which need further supervised fine-tune, here, we focus on providing representation that is only trained by unsupervised methods. This means models need to directly generate pixel-level, linearly separable semantic results. We first explore and present two factors that have significant effects on segmentation under the contrastive learning framework: 1) the difficulty and diversity of the positive contrastive pairs, 2) the balance of global and local features. With the intention of optimizing these factors, we propose the cycle-attention contrastive learning (CACL). CACL makes use of semantic continuity of video frames, adopting unsupervised cycle-consistent attention mechanism to implicitly conduct contrastive learning with difficult, global-local-balanced positive pixel pairs. Compared with baseline model MoCo-v2 and other unsupervised methods, CACL demonstrates consistently superior performance on PASCAL VOC (+4.5 mIoU) and Cityscapes (+4.5 mIoU) datasets.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

We consider the problem of learning models of stochastic environments for planning, assuming that the transitions on the various state factors are independent. We suppose we have example trajectories of policies for random problems executing in an unknown environment, and we would like to match the rate of success of those policies. We present a polynomial-time algorithm that, given a polynomial number of trajectories, produces a PPDDL model such that optimal policies under the PPDDL model will provably achieve a success rate at future goals that is almost as good or better than the policies that produced the training trajectories. Moreover, we also consider a variant of PPDDL in which there is uncertainty about the transition probabilities, specified by an interval for each factor that contains the respective true transition probabilities. We also give a polynomial-time algorithm that, when given the example trajectories, produces such an imprecise-PPDDL environment model and guarantees that with high probability, the true environment is indeed captured by the uncertain parameters, while also guaranteeing that the optimal bound on the success probability in this model is almost as good or better than the success rate of the policies that produced the training examples.

poster session 1 @ blue 4, poster session 11 @ blue 4, oral session 1 @ blue 4, poster session 1, poster session 11, oral session 1

In sequential decision making, objective specifications are often underspecified or incomplete, neglecting to take into account potential (negative) side effects. Executing plans without consideration of their side effects can lead to catastrophic outcomes -- a concern recently raised in relation to the safety of AI. In this paper we investigate how to avoid side effects in a symbolic planning setting. We study the notion of minimizing side effects in the context of a planning environment where multiple independent agents co-exist. We define (classes of) negative side effects in terms of their effect on the agency of those other agents. Finally, we show how plans which minimize side effects of different types can be computed via compilations to cost-optimizing symbolic planning, and investigate experimentally.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Temporal graphs naturally model graphs whose underlying topology changes over time. Recently, the problems Temporal Vertex Cover (or TVC) and Sliding-Window Temporal Vertex Cover (or $\Delta$-TVC for time-windows of a fixed-length $\Delta$) have been established as natural extensions of the classic Vertex Cover problem on static graphs with connections to areas such as surveillance in sensor networks. In this paper we initiate a systematic study of the complexity of TVC and $\Delta$-TVC on sparse graphs. Our main result shows that for every $\Delta\geq 2$, $\Delta$-TVC is NP-hard even when the underlying topology is described by a path or a cycle. This resolves an open problem from literature and shows a surprising contrast between $\Delta$-TVC and TVC for which we provide a polynomial-time algorithm in the same setting. To circumvent this hardness, we present a number of exact and approximation algorithms for temporal graphs whose underlying topologies are given by a path, that have bounded vertex degree in every time step, or that admit a small-sized temporal vertex cover.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Bugs in source files (SFs) may cause software malfunction, inconveniencing users and even leading to catastrophic accidents. Therefore, the bugs in SFs should be found and fixed quickly. However, from hundreds of candidate SFs, finding buggy SFs is tedious and time consuming. To lessen the burden on developers, deep learning-based bug localization (DLBL) tools can be utilized. Text terms in bug reports and SFs play an important role. However, some terms provide incorrect information and degrade bug localization performance. Therefore, those terms are defined here as "misguiding terms," and an explainable-artificial-intelligence-based identification method is proposed. The effectiveness of the proposed method for DLBL was investigated. When misguiding terms were removed, the mean average precision of the bug localization model improved by 33% on average.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

While commonsense knowledge acquisition and reasoning has traditionally been a core research topic in the knowledge representation and reasoning community, recent years have seen a surge of interest in the natural language processing community in developing pre-trained models and testing their ability to address a variety of newly designed commonsense knowledge reasoning and generation tasks. This paper presents a survey of these tasks, discusses the strengths and weaknesses of state-of-the-art pre-trained models for commonsense reasoning and generation as revealed by these tasks, and reflects on future research directions.

poster session 1 @ red 5, poster session 10 @ red 5, oral session 1 @ red 5, poster session 1, poster session 10, oral session 1

Offline Reinforcement Learning (RL) aims at learning an optimal control from a fixed dataset, without interactions with the system. An agent in this setting should avoid selecting actions whose consequences cannot be predicted from the data. This is the converse of exploration in RL, which favors such actions. We thus take inspiration from the literature on bonus-based exploration to design a new offline RL agent. The core idea is to subtract a prediction-based exploration bonus from the reward, instead of adding it for exploration. This allows the policy to stay close to the support of the dataset and practically extends some previous pessimism-based offline RL methods to a deep learning setting with arbitrary bonuses. We also connect this approach to a more common regularization of the learned policy towards the data. Instantiated with a bonus based on the prediction error of a variational autoencoder, we show that our simple agent is competitive with the state of the art on a set of continuous control locomotion and manipulation tasks.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

We study the problem of safe offline reinforcement learning (RL), the goal is to learn a policy that maximizes long-term reward while satisfying safety constraints given only offline data, without further interaction with the environment. This problem is more appealing for real world RL applications, in which data collection is costly or dangerous. Enforcing constraint satisfaction is non-trivial, especially in offline settings, as there is a potential large discrepancy between the policy distribution and the data distribution, causing errors in estimating the value of safety constraints. We show that na\"ive approaches that combine techniques from safe RL and offline RL can only learn sub-optimal solutions. We thus develop a simple yet effective algorithm, Constraints Penalized Q-Learning (CPQ), to solve the problem. Our method admits the use of data generated by mixed behavior policies. We present a theoretical analysis and demonstrate empirically that our approach can learn robustly across a variety of benchmark control tasks, outperforming several baselines.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

We present a novel non-iterative learnable method for partial-to-partial 3D shape registration. The partial alignment task is extremely complex, as it jointly tries to match between points, and identify which points do not appear in the corresponding shape, causing the solution to be non-unique and ill-posed in most cases. Until now, two main methodologies have been suggested to solve this problem: sample a subset of points that are likely to have correspondences, or perform soft alignment between the point clouds and try to avoid a match to an occluded part. These heuristics work when the partiality is mild or when the transformation is small but fails for severe occlusions, or when outliers are present. We present a novel approach named Confidence Guided Distance Network, where we fuse learnable similarity between point embeddings and spatial distance between point clouds, inducing an optimized solution for the overlapping points while ignoring parts that only appear in one of the shapes. The point feature generation is done by a self-supervised architecture that repels far points to have different embeddings, therefore succeeds to align partial views of shapes, even with excessive internal symmetries, or acute rotations. We compare our network to recently presented learning-based and axiomatic methods and report a fundamental boost in performance.

poster session 1 @ red 1, poster session 11 @ red 1, oral session 11 @ red 1, poster session 1, poster session 11, oral session 11

There is growing evidence that the classical notion of adversarial robustness originally introduced for images has been adopted as a de facto standard by a large part of the NLP research community. We show that this notion is problematic in the context of NLP as it considers a narrow spectrum of linguistic phenomena. In this paper, we argue for semantic robustness, which is better aligned with the human concept of linguistic fidelity. We characterize semantic robustness in terms of biases that it is expected to induce in a model. We study semantic robustness of a range of vanilla and robustly trained architectures using a template-based generative test bed. We complement the analysis with empirical evidence that, despite being harder to implement, semantic robustness can improve performance %gives guarantees for on complex linguistic phenomena where models robust in the classical sense fail.

poster session 1 @ red 4, poster session 8 @ red 4, oral session 1 @ red 4, poster session 1, poster session 8, oral session 1

Recent research on model interpretability in natural language processing extensively uses feature scoring methods for identifying which parts of the input are the most important for a model to make a prediction (i.e. explanation or rationale). However, previous research has shown that there is no clear best scoring method across various text classification tasks while practitioners typically have to make several other ad-hoc choices regarding the length and the type of the rationale (e.g. short or long, contiguous or not). Inspired by this, we propose a simple yet effective and flexible method that allows selecting optimally for each data instance: (1) a feature scoring method; (2) the length; and (3) the type of the rationale. Our method is inspired by input erasure approaches to interpretability which assume that the most faithful rationale for a prediction should be the one with the highest difference between the model's output distribution using the full text and the text after removing the rationale as input respectively. Evaluation on four standard text classification datasets shows that our proposed method provides more faithful, comprehensive and highly sufficient explanations compared to using a fixed feature scoring method, rationale length and type. More importantly, we demonstrate that a practitioner is not required to make any ad-hoc choices in order to extract faithful rationales using our approach.

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

The problem of knowledge-based visual question answering involves answering questions that require external knowledge in addition to the content of the image. Such knowledge typically comes in a variety of forms, including visual, textual, and commonsense knowledge. The use of more knowledge sources, however, also increases the chance of retrieving more irrelevant or noisy facts, making it difficult to comprehend the facts and find the answer. To address this challenge, we propose Multi-modal Answer Validation using External knowledge (MAVEx), where the idea is to validate a set of promising answer candidates based on answer-specific knowledge retrieval. This is in contrast to existing approaches that search for the answer in a vast collection of often irrelevant facts. Our approach aims to learn which knowledge source should be trusted for each answer candidate and how to validate the candidate using that source. We consider a multi-modal setting, relying on both textual and visual knowledge resources, including images searched using Google, sentences from Wikipedia articles, and concepts from ConceptNet. Our experiments with OK-VQA, a challenging knowledge-based VQA dataset, demonstrate that \mavex achieves new state-of-the-art results.

poster session 1 @ blue 4, poster session 11 @ blue 4, oral session 11 @ blue 4, poster session 1, poster session 11, oral session 11

Graph walking based on reinforcement learning (RL) has shown great success in navigating an agent to automatically complete various reasoning tasks over an incomplete knowledge graph (KG) by exploring multi-hop relational paths. However, existing multi-hop reasoning approaches only work well on short reasoning paths and tend to miss the target entity with the increasing path length. This is undesirable for many reasoning tasks in real-world scenarios, where short paths connecting the source and target entities are not available in incomplete KGs, and thus the reasoning performances drop drastically unless the agent is able to seek out more clues from longer paths. To address the above challenge, in this paper, we propose a dual-agent reinforcement learning framework, which trains two agents (\ts{giant} and \ts{dwarf}) to walk over a KG \ti{jointly} and search for the answer \ti{collaboratively}. Our approach tackles the reasoning challenge in long paths by assigning one of the agents (\ts{giant}) searching on cluster-level paths quickly and providing stage-wise hints for another agent (\ts{dwarf}). Finally, experimental results on several KG reasoning benchmarks show that our approach can search answers more accurately and efficiently, and outperforms existing RL-based methods for long path queries by a large margin.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

Recent self-supervised video representation learning methods have found significant success by exploring essential properties of videos, e.g. speed, temporal order, etc. This work exploits an essential yet under-explored property of videos, the \textit{video continuity}, to obtain supervision signals in self-supervised representation learning. Specifically, we formulate three novel continuity-related pretext tasks, i.e. continuity justification, discontinuity localization, and missing section approximation, that supervise a shared backbone for video representation learning. This self-supervision approach, termed as Continuity Perception Network (CPNet), encourages its backbone network to learn local and long-ranged motion and context representations and outperforms prior arts on multiple downstream tasks, such as action recognition, video retrieval, and action localization. Also, the video continuity can be complementary to other video properties for representation learning, and integrating the proposed pretext task to prior arts can yield much performance gains.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

In recent years, the automation of machine learning and data science (AutoML) has attracted significant attention. One under-explored dimension of AutoML is being able to automatically utilize domain knowledge (such as semantic concepts and relationships) located in historical code or literature from the problem's domain. In this paper, we demonstrate Semantic Feature Discovery, which enables users to interactively explore features semantically discovered from existing data science code and external knowledge. It does so by detecting semantic concepts for a given dataset, and then using these concepts to determine relevant feature engineering operations from historical code and knowledge.

poster session 1 @ red 5, poster session 10 @ red 5, poster session 1, poster session 10

Online travel platforms (OTPs), e.g., bookings.com and Ctrip.com, deliver travel experiences to online users by providing travel-related products. Although much progress has been made, the state-of-the-arts for cold-start problems are largely sub-optimal for user representation, since they do not take into account the unique characteristics exhibited from the user behaviors. In this work, we propose a State-aware Multi-aspect Interests representation Network (SMINet) for cold-start users recommendation at OTPs, which consists of a multi-aspect interests extractor, a co-attention layer, and a state-aware gating layer. The key component of the model is the multi-aspect interests extractor, which is able to extract representations for the user's multi-aspect interests.Furthermore, to learn the interactions between the user behaviors in the current session and the above multi-aspect interests, we carefully design a co-attention layer which allows the cross attentions between the two modules. Additionally, we propose a travel state-aware gating layer to attentively select the multi-aspect interests. The final user representation is obtained by fusing the three components. Comprehensive experiments conducted both offline and online demonstrate the superior performance of the proposed model at user representation, especially for cold-start users, compared with state-of-the-art methods.

poster session 1 @ blue 5, poster session 11 @ blue 5, oral session 11 @ blue 5, poster session 1, poster session 11, oral session 11

From ecology to atmospheric sciences, many academic disciplines deal with data characterized by intricate spatio-temporal complexities, the modeling of which often requires specialized approaches. Generative models of these data are of particular interest, as they enable a range of impactful downstream applications like simulation or creating synthetic training data. Recently, COT-GAN, a new GAN algorithm inspired by the theory of causal optimal transport (COT), was proposed in an attempt to improve generation of sequential data. However, the task of learning complex patterns over time and space requires additional knowledge of the specific data structures. In this study, we propose a novel loss objective combined with COT-GAN based on an autoregressive embedding to reinforce the learning of spatio-temporal dynamics. We devise SPATE (spatio-temporal association), a new metric measuring spatio-temporal autocorrelation. We compute SPATE for real and synthetic data samples and use it to compute an embedding loss that considers space-time interactions, nudging the GAN to learn outputs that are faithful to the observed dynamics. We test our new SPATE-GAN on a diverse set of spatio-temporal patterns: turbulent flows, log-Gaussian Cox processes and global weather data. We show that our novel embedding loss improves performance without any changes to the architecture of the GAN backbone, highlighting our model's increased capacity for capturing autoregressive structures.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

Recent studies indicate that deep neural networks (DNNs) are prone to show discrimination towards certain demographic groups. We observe that algorithmic discrimination can be explained by the high reliance of the models on fairness sensitive features. Motivated by this observation, we propose to achieve fairness by suppressing the DNN models from capturing the spurious correlation between those fairness sensitive features with the underlying task. Specifically, we firstly train a bias-only teacher model which is explicitly encouraged to maximally employ fairness sensitive features for prediction. The teacher model then counter-teaches a debiased student model so that the interpretation of the student model is orthogonal to the interpretation of the teacher model. The key idea is that since the teacher model relies explicitly on fairness sensitive features for prediction, the orthogonal interpretation loss enforces the student network to reduce its reliance on sensitive features and instead capture more task relevant features for prediction. Experimental analysis indicates that our framework substantially reduces the model's attention on fairness sensitive features. Experimental results on four datasets further validate that our framework has consistently improved the fairness with respect to three group fairness metrics, with a comparable or even better accuracy.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

Readers of novels need to identify and learn about the characters as they develop an understanding of the plot. The paper presents an end-to-end automated pipeline for literary character identification and ongoing work for extracting and comparing character representations for full-length English novels. The character identification pipeline involves a named entity recognition (NER) module with F1 score of 0.85, a coreference resolution module with F1 score of 0.76, and a disambiguation module using both heuristic and algorithmic approaches. Ongoing work compares event extraction as well as speech extraction pipelines for literary characters representations with case studies. The paper is the first to my knowledge that combines a modular pipeline for automated character identification, representation extraction and comparisons for full-length English novels.

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

Spiking neural networks (SNNs) can be run on neuromorphic devices with ultra-high speed and ultra-low energy consumption because of their binary and event-driven nature. Therefore, SNNs are expected to have various applications, including as generative models being running on edge devices to create high-quality images. In this study, we build a variational autoencoder (VAE) with SNN to enable image generation. VAE is known for its stability among generative models; recently, its quality advanced. In vanilla VAE, the latent space is represented as a normal distribution, and floating-point calculations are required in sampling. However, this is not possible in SNNs because all features must be binary time series data. Therefore, we constructed the latent space with an autoregressive SNN model, and randomly selected samples from its output to sample the latent variables. This allows the latent variables to follow the Bernoulli process and allows variational learning. Thus, we build the Fully Spiking Variational Autoencoder where all modules are constructed with SNN. To the best of our knowledge, we are the first to build a VAE only with SNN layers. We experimented with several datasets, and confirmed that it can generate images with the same or better quality compared to conventional ANNs.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Given the incomplete knowledge of classes that exist in the world, Open-set Recognition (OSR) enables networks to identify and reject the unseen classes after training. This problem of breaking the common closed-set assumption is far from being solved. Recent studies focus on designing new losses, neural network encoding structures, and calibration methods to optimize a feature space for OSR relevant tasks. In this work, we make the first attempt to tackle OSR by searching the architecture of a Neural Network (NN) under the open-set assumption. In contrast to the prior arts, we develop a mechanism to both search the architecture of the network and train a network suitable for tackling OSR. Inspired by the compact abating probability (CAP) model, which is theoretically proven to reduce the open space risk, we regularize the searching space by VAE contrastive learning. To discover a more robust structure for OSR, we propose Pseudo Auxiliary Searching (PAS), in which we split a pretended set of know-unknown classes from the original training set in the searching phase, hence enabling the super-net to explore an effective architecture that can handle unseen classes in advance. We demonstrate the benefits of this learning pipeline on 5 OSR datasets, including MNIST, SVHN, CIFAR10, CIFARAdd10, and CIFARAdd50, where our approach outperforms prior state-of-the-art networks designed by humans.

poster session 1 @ red 3, poster session 11 @ red 3, oral session 1 @ red 3, poster session 1, poster session 11, oral session 1

Broadly speaking, I hope to dedicate my PhD to improving our understanding of algorithmic economics with the ultimate goal of building welfare improving decentralized technology for markets. In the following pages, I describe how my past work has built on the existing literature to get closer to the goal of creating such technologies, and describe what research paths this work opens up for the rest of my PhD. I believe that my research has the potential to provide algorithmic solutions to problems in machine learning, optimization, and game theory, and can be used to improve the efficiency of online marketplaces.

poster session 1 @ blue 6, poster session 2 @ blue 6, poster session 1, poster session 2

Data in the real world tends to exhibit a long-tailed label distribution, which poses great challenges for the training of neural networks in visual recognition. Existing methods tackle this problem mainly from the perspective of data quantity, i.e., the number of samples in each class. To be specific, they pay more attention to tail classes, like applying larger adjustments to the logit. However, in the training process, the quantity and difficulty of data are two intertwined and equally crucial problems. For some tail classes, the features of their instances are distinct and discriminative, which can also bring satisfactory accuracy; for some head classes, although with sufficient samples, the high semantic similarity with other classes and lack of discriminative features will bring bad accuracy. Based on these observations, we propose Adaptive Logit Adjustment Loss (ALA Loss) to apply an adaptive adjusting term to the logit. The adaptive adjusting term is composed of two complementary factors: 1) quantity factor, which pays more attention to tail classes, and 2) difficulty factor, which adaptively pays more attention to hard instances in the training process. The difficulty factor can alleviate the over-optimization on tail yet easy instances and under-optimization on head yet hard instances. The synergy of the two factors can not only advance the performance on tail classes even further, but also promote the accuracy on head classes. Unlike previous logit adjusting methods that only concerned about data quantity, ALA Loss tackles the long-tailed problem from a more comprehensive, fine-grained and adaptive perspective. Extensive experimental results show that our method achieves the state-of-the-art performance on challenging recognition benchmarks, including ImageNet-LT, iNaturalist 2018, and Places-LT.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

We explore the utility of information contained within a dropout based Bayesian neural network (BNN) for the task of detecting out of distribution (OOD) data. We first show how previous attempts to leverage the randomized embeddings induced by the intermediate layers of a dropout BNN can fail due to the distance metric used. We introduce an alternative approach to measuring embedding uncertainty, and demonstrate how incorporating embedding uncertainty improves OOD data identification across three tasks: image classification, language classification, and malware detection.

poster session 1 @ blue 5, poster session 11 @ blue 5, oral session 1 @ blue 5, poster session 1, poster session 11, oral session 1

Although considerable progress has been achieved regarding the transformers in recent years, the large number of parameters, quadratic computational complexity and memory cost conditioned on long sequences make the transformers hard to train and implement, especially in edge computing configurations. In this case, a dizzying number of works have sought to make improvements around computational and memory efficiency upon the original transformer architecture. Nevertheless, many of them restrict the context in the attention to seek a trade-off between cost and performance with prior knowledge of orderly stored data. It is imperative to dig deep into an efficient feature extractor for point clouds due to their irregularity and large quantity of points. In this paper, we propose a novel skeleton decomposition-based self-attention (SD-SA) which has no sequence length limit and exhibits favorable scalability in long-sequence models. Specifically, due to the numerical low-rank nature of self-attention, we approximate it by the skeleton decomposition method while maintaining its effectiveness. At this point, we have shown that the proposed method works for the proposed approach on point cloud classification, segmentation and detection tasks on the ModelNet40, ShapeNet, and KITTI dataset, respectively. Our approach significantly improves the efficiency of the point cloud transformer and exceeds other efficient transformers on point cloud tasks in terms of speed at comparable performance.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

In attempts to "explain" predictions of machine learning models, researchers have proposed hundreds of techniques for attributing predictions to features that are deemed important. While these attributions are often claimed to hold the potential to improve human "understanding" of the models, surprisingly little work explicitly evaluates progress towards this aspiration. In this paper, we conduct a crowdsourcing study, where participants interact with deception detection models that have been trained to distinguish between genuine and fake hotel reviews. They are challenged both to simulate the model on fresh reviews, and to edit reviews with the goal of lowering the probability of the originally predicted class. Successful manipulations would lead to an adversarial example. During the training (but not the test) phase, input spans are highlighted to communicate salience. Through our evaluation, we observe that for a linear bag-of-words model, participants with access to the feature coefficients during training are able to cause a larger reduction in model confidence in the testing phase when compared to the no-explanation control. For the BERT-based classifier, popular local explanations do not improve their ability to reduce the model confidence over the no-explanation case. Remarkably, when the explanation for the BERT model is given by the (global) attributions of a linear model trained to imitate the BERT model, people can effectively manipulate the model.

poster session 1 @ red 4, poster session 8 @ red 4, oral session 1 @ red 4, poster session 1, poster session 8, oral session 1

In many real-world scenarios, we often deal with streaming data that is sequentially collected over time. Due to the non-stationary nature of the environment, the streaming data distribution may change in unpredictable ways, which is known as the concept drift in the literature. To handle concept drift, previous methods first detect when/where the concept drift happens and then adapt models to fit the distribution of the latest data. However, there are still many cases that some underlying factors of environment evolution are predictable, making it possible to model the future concept drift trend of the streaming data, while such cases are not fully explored in previous work. In this paper, we propose a novel method DDG-DA, that can effectively forecast the evolution of data distribution and improve the performance of models. Specifically, we first train a predictor to estimate the future data distribution, then leverage it to generate training samples, and finally train models on the generated data. We conduct experiments on three real-world tasks (forecasting on stock price trend, electricity load and solar irradiance) and obtained significant improvement on multiple widely-used models.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

A promising trend in deep learning replaces traditional feedforward networks with implicit networks. Unlike traditional networks, implicit networks solve a fixed point equation to compute inferences. Solving for the fixed point varies in complexity, depending on provided data and an error tolerance. Importantly, implicit networks may be trained with fixed memory costs in stark contrast to feedforward networks, whose memory requirements scale linearly with depth. However, there is no free lunch --- backpropagation through implicit networks often requires solving a costly Jacobian-based equation arising from the implicit function theorem. We propose Jacobian-Free Backpropagation (JFB), a fixed-memory approach that circumvents the need to solve Jacobian-based equations. JFB makes implicit networks faster to train and significantly easier to implement, without sacrificing test accuracy. Our experiments show implicit networks trained with JFB are competitive with feedforward networks and prior implicit networks given the same number of parameters.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Voxel and point representations are widely applied in recent 3D object detection tasks from LiDAR point clouds. Voxel representations contribute to efficiently and rapidly locating objects, whereas point representations are capable of describing intra-object spatial relationship for detection refinement. In this work, we aim to exploit the strengths of both two representations, and present a novel two-stage detector, named Joint Point-Voxel Network (JPV-Net). Specifically, our framework is equipped with a Dual Encoders-Fusion Decoder, which consists of the dual encoders to extract voxel features of sketchy 3D scenes and point features rich in geometric context, respectively, and the Feature Propagation Fusion (FP-Fusion) decoder to attentively fuse them from coarse to fine. By making use of the advantages of these features, the refinement network can effectively eliminate false detection and provide better accuracy. Besides, to further develop the perception characteristics of voxel CNN and point backbone, we design two novel intersection-over-union (IoU) estimation modules for proposal generation and refinement, both of which can alleviate the misalignment between the localization and the classification confidence. Extensive experiments on the KITTI dataset and ONCE dataset demonstrate that our proposed JPV-Net outperforms other state-of-the-art methods with remarkable margins.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

Previous deep learning-based line segment detection (LSD) suffers from the immense model size and high computational cost for line prediction. This constrains them from real-time inference on computationally restricted environments. In this paper, we propose a real-time and light-weight line segment detector for resource-constrained environments named Mobile LSD (M-LSD). We design an extremely efficient LSD architecture by minimizing the backbone network and removing the typical multi-module process for line prediction found in previous methods. To maintain competitive performance with a light-weight network, we present novel training schemes: Segments of Line segment (SoL) augmentation, matching and geometric loss. SoL augmentation splits a line segment into multiple subparts, which are used to provide auxiliary line data during the training process. Moreover, the matching and geometric loss allow a model to capture additional geometric cues. Compared with TP-LSD-Lite, previously the best real-time LSD method, our model (M-LSD-tiny) achieves competitive performance with 2.5% of model size and an increase of 130.5% in inference speed on GPU. Furthermore, our model runs at 56.8 FPS and 48.6 FPS on the latest Android and iPhone mobile devices, respectively. To the best of our knowledge, this is the first real-time deep LSD available on mobile devices. Our code is available.

poster session 1 @ red 1, poster session 11 @ red 1, oral session 11 @ red 1, poster session 1, poster session 11, oral session 11

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

In this paper, we introduce VACA, a novel class of variational graph autoencoders for causal inference in the absence of hidden confounders, when only observational data and the causal graph are available. Without making any parametric assumptions, VACA mimics the necessary properties of a Structural Causal Model (SCM) to provide a flexible and practical framework for approximating interventions (do-operator) and abduction-action-prediction steps. As a result, and as shown by our empirical results, VACA accurately approximates the interventional and counterfactual distributions on diverse SCMs. Finally, we apply VACA to evaluate counterfactual fairness in fair classification problems, as well as to learn fair classifiers without compromising performance.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

Learning to understand grounded language, which connects natural language to percepts, is a critical research area. Prior work in grounded language acquisition has focused primarily on textual inputs. In this work we demonstrate the feasibility of performing grounded language acquisition on paired visual percepts and raw speech inputs. This will allow human-robot interactions in which language about novel tasks and environments is learned from end users, reducing dependence on textual inputs and potentially mitigating the effects of demographic bias found in widely available speech recognition systems. We leverage recent work in self-supervised speech representation models and show that learned representations of speech can make language grounding systems more inclusive towards specific groups while maintaining or even increasing general performance.

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

Aspect Based Sentiment Analysis (ABSA) aims to extract aspect terms and identify the sentiment polarities towards each extracted aspect term. Currently, syntactic information is seen as the bridge for the domain adaptation and achieves remarkable performance. However, the transferable syntactic knowledge is complex and diverse, which causes the transfer error problem in domain adaptation. In our paper, we propose a domain-shared relational structure incorporated cross-domain ABSA model. The experimental results show the effectiveness of our model.

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

Object localization aims to generate a tight bounding box for the target object, which is a challenging problem that has been deeply studied in recent years. Since collecting bounding-box labels is time-consuming and laborious, many researchers focus on weakly supervised object localization (WSOL). As the recent appealing self-supervised learning technique shows its powerful function in visual tasks, in this paper, we take the early attempt to explore unsupervised object localization by self-supervision. Specifically, we adopt different geometric transformations to image and utilize their parameters as pseudo labels for self-supervised learning. Then, the class-agnostic activation map (CAAM) is used to highlight the target object potential regions. However, such attention maps merely focus on the most discriminative part of the objects, which will affect the quality of the predicted bounding box. Based on the motivation that the activation maps of different transformations of the same image should be equivariant, we further design a siamese network that encodes the paired images and propose a joint graph cluster partition mechanism in an unsupervised manner to enhance the object co-occurrent regions. To validate the effectiveness of the proposed method, extensive experiments are conducted on CUB-200-2011, Stanford Cars and FGVC-Aircraft datasets. Experimental results show that our method outperforms state-of-the-art methods using the same level of supervision, even outperforms some weakly-supervised methods.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

Long Short-Term Memory (LSTM) and Transformers are two popular neural architectures used for natural language processing tasks. Theoretical results show that both are Turing-complete and can represent any context-free language (CFL).In practice, it is often observed that Transformer models have better representation power than LSTM. But the reason is barely understood. We study such practical differences between LSTM and Transformer and propose an explanation based on their latent space decomposition patterns. To achieve this goal, we introduce an oracle training paradigm, which forces the decomposition of the latent representation of LSTMand the Transformer and supervises with the transitions of the Pushdown Automaton (PDA) of the corresponding CFL. With the forced decomposition, we show that the performance upper bounds of LSTM and Transformer in learning CFL are close: both of them can simulate a stack and perform stack operation along with state transitions. However, the absence of forced decomposition leads to the failure of LSTM models to capture the stack and stack operations, while having a marginal impact on the Transformer model. Lastly, we connect the experiment on the prototypical PDA to a real-world parsing task to re-verify the conclusions

poster session 1 @ blue 2, poster session 11 @ blue 2, oral session 1 @ blue 2, poster session 1, poster session 11, oral session 1

Goal recognition design (GRD) is the task of modifying environments for aiding observers to recognize the objectives of agents during online observations. The worst case distinctiveness (WCD), a widely used performance measure in GRD research, can fail to provide useful guidance to the redesign process when some goals are too hard to be distinguished. Moreover, the existing WCD-based approaches do not work when an agent aims for a sequence of goals instead of just one goal. The paper presents a new GRD framework called extended goal recognition design (EGRD) for goal recognition that involves multiple goals. The objective of EGRD is to modify an environment to minimize the worst case distinctiveness of a goal condition that describes how an agent can reach a set of goals. A goal condition can be formally expressed in first-order computation tree logic (FO-CTL) that can be evaluated by model checking. We introduce a novel graphical representation of FO-CTL sentences that is suitable for extended goal recognition. Moreover, we present a search algorithm for EGRD with a novel caching mechanism. Our experimental results show that the caching mechanism can greatly speed up our EGRD search algorithm by reusing the previous evaluation of FO-CTL sentences.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Traditional Chinese Medicine (TCM) constitution is a fundamental concept in TCM theory. It is determined by multimodal TCM clinical features which, in turn, are obtained from TCM clinical information of image (face, tongue, etc.), audio (pulse and voice), and text (inquiry) modality. The auto assessment of TCM constitution is faced with two major challenges: (1) learning discriminative TCM clinical feature representations; (2) jointly processing the features using multimodal fusion techniques. The TCM Constitution Assessment System (TCM-CAS) is proposed to provide an end-to-end solution to this task, along with auxiliary functions to aid TCM researchers. To improve the results of TCM constitution prediction, the system combines multiple machine learning algorithms such as facial landmark detection, image segmentation, graph neural networks and multimodal fusion. Extensive experiments are conducted on a four-category multimodal TCM constitution dataset, and the proposed method achieves state-of-the-art accuracy. Provided with datasets containing annotations of diseases, the system can also perform automatic disease diagnosis from a TCM perspective.

poster session 1 @ red 5, poster session 10 @ red 5, poster session 1, poster session 10

Neural ordinary differential equations (NODE) have been proposed as a continuous depth generalization to popular deep learning models such as Residual networks (ResNets). They provide parameter efficiency and automate the model selection process in deep learning models to some extent. However, they lack the much-required uncertainty modelling and robustness capabilities which are crucial for their use in several real-world applications such as autonomous driving and healthcare. We propose a novel and unique approach to model uncertainty in NODE by considering a distribution over the end-time T of the ODE solver. The proposed approach, latent time NODE (LT-NODE), treats T as a latent variable and apply Bayesian learning to obtain a posterior distribution over T from the data. In particular, we use variational inference to learn an approximate posterior and the model parameters. Prediction is done by considering the NODE representations from different samples of the posterior and can be done efficiently using a single forward pass. As T implicitly defines the depth of a NODE, posterior distribution over T would also help in model selection in NODE. We also propose, adaptive latent time NODE (ALT-NODE), which allow each data point to have a distinct posterior distribution over end-times. ALT-NODE uses amortized variational inference to learn an approximate posterior using inference networks. We demonstrate the effectiveness of the proposed approaches in modelling uncertainty and robustness through experiments on synthetic and several real-world image classification data.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

Inspired by the extensive success of deep learning, graph neural networks (GNNs) have been proposed to learn expressive node representations and demonstrated promising performance in various graph learning tasks. However, existing endeavors predominately focus on the conventional semi-supervised setting where relatively abundant gold-labeled nodes are provided. While it is often impractical due to the fact that data labeling is unbearably laborious and requires intensive domain knowledge, especially when considering the heterogeneity of graph-structured data. Under the few-shot semi-supervised setting, the performance of most of the existing GNNs is inevitably undermined by the overfitting and oversmoothing issues, largely owing to the shortage of labeled data. In this paper, we propose a decoupled network architecture equipped with a novel meta-learning algorithm to solve this problem. In essence, our framework Meta-PN infers high-quality pseudo labels on unlabeled nodes via a meta-learned label propagation strategy, which effectively augments the scarce labeled data while enabling large receptive fields during training. Extensive experiments demonstrate that our approach offers easy and substantial performance gains compared to existing techniques on various benchmark datasets.

poster session 1 @ blue 1, poster session 8 @ blue 1, oral session 8 @ blue 1, poster session 1, poster session 8, oral session 8

We investigate the fairness concerns of training a machine learning model using data with missing values. Even though there are a number of fairness intervention methods in the literature, most of them require a complete training set as input. In practice, data can have missing values, and data missing patterns can depend on group attributes (e.g. gender or race). Simply applying off-the-shelf fair learning algorithms to an imputed dataset may lead to an unfair model. In this paper, we first theoretically analyze different sources of discrimination risks when training with an imputed dataset. Then, we propose an integrated approach based on decision trees that does not require a separate process of imputation and learning. Instead, we train a tree with missing incorporated as attribute (MIA), which does not require explicit imputation, and we optimize a fairness-regularized objective function. We demonstrate that our approach outperforms existing fairness intervention methods applied to an imputed dataset, through several experiments on real-world datasets.

poster session 1 @ blue 3, poster session 12 @ blue 3, oral session 12 @ blue 3, poster session 1, poster session 12, oral session 12

Existing facial analysis systems have been shown to yield biased results against certain demographic subgroups. Due to its impact on society, it has become imperative to ensure that these systems do not discriminate based on gender, identity, or skin tone of individuals. This has led to research in the identification and mitigation of bias in AI systems. In this paper, we encapsulate bias detection/estimation and mitigation algorithms for facial analysis. Our main contributions include a systematic review of algorithms proposed for understanding bias, along with a taxonomy and extensive overview of existing bias mitigation algorithms. We also discuss open challenges in the field of biased facial analysis.

poster session 1 @ red 5, poster session 10 @ red 5, oral session 1 @ red 5, poster session 1, poster session 10, oral session 1

A recent report of Littmann [Commun. ACM '21] outlines the existence and the fatal impact of collusion rings in academic peer reviewing. We introduce and analyze the problem Cycle-Free Reviewing that aims at finding a review assignment without the following kind of collusion ring: A sequence of reviewers each reviewing a paper authored by the next reviewer in the sequence (with the last reviewer reviewing a paper of the first), thus creating a review cycle where each reviewer gives favorable reviews. As a result, all papers in that cycle have a high chance of acceptance independent of their respective scientific merit. We observe that review assignments computed using a standard Linear Programming approach typically admit many short review cycles. On the negative side, we show that Cycle-Free Reviewing is NP-hard in various restricted cases (i.e., when every author is qualified to review all papers and one wants to prevent that authors review each other's or their own papers or when every author has only one paper and is only qualified to review few papers). On the positive side, among others, we show that, in some realistic settings, an assignment without any review cycles of small length always exists. This result also gives rise to an efficient heuristic for computing (weighted) cycle-free review assignments, which we show to be of excellent quality in practice.

poster session 1 @ blue 6, poster session 12 @ blue 6, poster session 1, poster session 12

In this work, we present SenTag, a lightweight web-based tool focused on semantic annotation of textual documents. The platform allows multiple users to work on a corpus of documents. The tool enables to tag a corpus of documents through an intuitive and easy-to-use user interface that adopts the Extensible Markup Language (XML) as output format. The main goal of the application is two-fold: facilitating the tagging process and reducing or avoiding errors in the output documents. It allows also to identify arguments and other entities that are used to build an arguments graph. It is also possible to assess the level of agreement of annotators working on a corpus of text.

poster session 1 @ red 5, poster session 10 @ red 5, poster session 1, poster session 10

Variational Autoencoders (VAEs) suffer from degenerated performance, when learning several successive tasks. This is caused by catastrophic forgetting. In order to address the knowledge loss, VAEs are using either Generative Replay (GR) mechanisms or Expanding Network Architectures (ENA). In this paper we study the forgetting behaviour of VAEs using a joint GR and ENA methodology, by deriving an upper bound on the negative marginal log-likelihood. This theoretical analysis provides new insights into how VAEs forget the previously learnt knowledge during lifelong learning. The analysis indicates the best performance achieved when considering model mixtures, under the ENA framework, where there are no restrictions on the number of components. However, an ENA-based approach may require an excessive number of parameters. This motivates us to propose a novel Dynamic Expansion Graph Model (DEGM). DEGM expands its architecture, according to the novelty associated with each new databases, when compared to the information already learnt by the network from previous tasks. DEGM training optimizes knowledge structuring, characterizing the joint probabilistic representations corresponding to the past and more recently learned tasks. We demonstrate that DEGM guarantees optimal performance for each task while also minimizing the required number of parameters. Supplementary materials (SM) and source code are available in https://github.com/dtuzi123/Expansion-Graph-Model.

poster session 1 @ blue 2, poster session 11 @ blue 2, oral session 11 @ blue 2, poster session 1, poster session 11, oral session 11

Spatiotemporal modeling in an unified architecture is key for video action recognition. This paper proposes a Shrinking Temporal Attention Transformer (STAT), which efficiently builts spatiotemporal attention maps considering the attenuation of spatial attention in short and long temporal sequences. Specifically, for short-term temporal tokens, query token interacts with them in a fine-grained manner in dealing with short-range motion. It then shrinks to a coarse attention in neighborhood for long-term tokens, to provide larger receptive field for long-range spatial aggregation. Both of them are composed in a short-long temporal integrated block to build visual appearances and temporal structure concurrently with lower costly in computation. We conduct thorough ablation studies, and achieve state-of-the-art results on multiple action recognition benchmarks including Kinetics400 and Something-Something v2, outperforming prior methods with 50% less FLOPs and without any pretrained model.

poster session 1 @ red 1, poster session 11 @ red 1, poster session 1, poster session 11

The entertainment industry, as well as the field of Computer Graphics, frequently faces the issue of creating large virtual crowds that would populate a scene. One of the ways to achieve that, particularly with modern rendering techniques, is by using simulation -- this, however, is nontrivial to design and control. The main goal of my PhD work is working towards the creation of a tool enabling the creation of virtual crowds that one can interact with, and we believe the best way to that is through Multiagent Reinforcement Learning techniques. These animated crowds can then be used both in movies and video games. Especially for the latter, it is highly desirable that both the crowd as a whole, as well as the individual characters, can react to the user's input in real time.

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

Simplicial neural networks (SNNs) have recently emerged as a new direction in graph learning which expands the idea of convolutional architectures from node space to simplicial complexes on graphs. Instead of predominantly assessing pairwise relations among nodes as in the current practice, simplicial complexes allow us to describe higher-order interactions and multi-node graph structures. By building upon connection between the convolution operation and the new block Hodge-Laplacian, we propose the first SNN for link prediction. Our new Block Simplicial Complex Neural Networks (BScNets) model generalizes existing graph convolutional network (GCN) frameworks by systematically incorporating salient interactions among multiple higher-order graph structures of different dimensions. We discuss theoretical foundations behind BScNets and illustrate its utility for link prediction on eight real-world and synthetic datasets. Our experiments indicate that BScNets outperforms the state-of-the-art models by a significant margin while maintaining low computation costs. Finally, we show utility of BScNets as a new promising alternative for tracking spread of infectious diseases such as COVID-19 and measuring the effectiveness of the healthcare risk mitigation strategies.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

Multimodal sensors (visual, non-visual, and wearable) can provide complementary information to develop robust perception systems for recognizing activities accurately. However, it is challenging to extract robust multimodal representations due to the heterogeneous characteristics of data from multimodal sensors and disparate human activities, especially in the presence of noisy and misaligned sensor data. In this work, we propose a cooperative multitask learning-based guided multimodal fusion approach, MuMu, to extract robust multimodal representations for human activity recognition (HAR). MuMu employs an auxiliary task learning approach to extract features specific to each set of activities with shared characteristics (activity-group). MuMu then utilizes activity-group-specific features to direct our proposed Guided Multimodal Fusion Approach (GM-Fusion) for extracting complementary multimodal representations, designed as the target task. We evaluated MuMu by comparing its performance to state-of-the-art multimodal HAR approaches on three activity datasets. Our extensive experimental results suggest that MuMu outperforms all the evaluated approaches across all three datasets. Additionally, the ablation study suggests that MuMu significantly outperforms the baseline models (p<0.05), which do not use our guided multimodal fusion. Finally, the robust performance of MuMu on noisy and misaligned sensor data posits that our approach is suitable for HAR in real-world settings.

poster session 1 @ red 1, poster session 11 @ red 1, oral session 1 @ red 1, poster session 1, poster session 11, oral session 1

In recent years, plentiful evidence illustrates that Graph Convolutional Networks (GCNs) achieve extraordinary accomplishments on the node classification task. However, GCNs may be vulnerable to adversarial attacks on label-scarce dynamic graphs. Many existing works aim to strengthen the robustness of GCNs; for instance, adversarial training is used to shield GCNs against malicious perturbations. However, these works fail on dynamic graphs for which label scarcity is a pressing issue. To overcome label scarcity, self-training attempts to iteratively assign pseudo-labels to highly confident unlabeled nodes but such attempts may suffer serious degradation under dynamic graph perturbations. In this paper, we generalize noisy supervision as a kind of self-supervised learning method and then propose a novel Bayesian self-supervision model, namely GraphSS, to address the issue. Extensive experiments demonstrate that GraphSS can not only affirmatively alert the perturbations on dynamic graphs but also effectively recover the prediction of a node classifier when the graph is under such perturbations. These two advantages prove to be generalized over three classic GCNs across five public graph datasets.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

When applying multi-instance learning (MIL) to make predictions for bags of instances, the prediction accuracy of an instance often depends on not only the instance itself but also its context in the corresponding bag. From the viewpoint of causal inference, such bag contextual prior works as a confounder and may result in model robustness and interpretability issues. Focusing on this problem, we propose a novel interventional multi-instance learning (IMIL) framework to achieve deconfounded instance-level prediction. Unlike traditional likelihood-based strategies, we design an Expectation-Maximization (EM) algorithm based on causal intervention, providing a robust instance selection in the training phase and suppressing the bias caused by the bag contextual prior. Experiments on pathological image analysis demonstrate that our IMIL method substantially reduces false positives and outperforms state-of-the-art MIL methods.

poster session 1 @ red 2, poster session 10 @ red 2, poster session 1, poster session 10

The recently proposed DEtection TRansformer (DETR) achieves promising performance for end-to-end object detection. However, it has relatively lower detection performance on small objects and suffers from slow convergence. This paper observed that DETR performs surprisingly well even on small objects when measuring Average Precision (AP) at decreased Intersection-over-Union (IoU) thresholds. Motivated by this observation, we propose a simple way to improve DETR by refining the coarse features and predicted locations. Specifically, we propose a novel Coarse-to-Fine (CF) decoder layer constituted of a coarse layer and a carefully designed fine layer. Within each CF decoder layer, the extracted local information (region of interest feature) is introduced into the flow of global context information from the coarse layer to refine and enrich the object query features via the fine layer. In the fine layer, the multi-scale information can be fully explored and exploited via the Adaptive Scale Fusion(ASF) module and Local Cross-Attention (LCA) module. The multi-scale information can also be enhanced by another proposed Transformer Enhanced FPN (TEF) module to further improve the performance. With our proposed framework (named CF-DETR), the localization accuracy of objects (especially for small objects) can be largely improved. As a byproduct, the slow convergence issue of DETR can also be addressed. The effectiveness of CF-DETR is validated via extensive experiments on the coco benchmark. CF-DETR achieves state-of-the-art performance among end-to-end detectors, e.g., achieving 47.8 AP using ResNet-50 with 36 epochs in the standard 3x training schedule.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11

This work aims at improving instance retrieval with self-supervision. We find that fine-tuning using the recently developed self-supervised (SSL) learning methods, such as SimCLR and MoCo, fails to improve the performance of instance retrieval. In this work, we identify that the learnt representations for instance retrieval should be invariant to large variations in viewpoint and background etc., whereas self-augmented positives applied by the current SSL methods can not provide strong enough signals for learning robust instance-level representations. To overcome this problem, we propose InsCLR, a new SSL method that builds on the instance-level contrast, to learn the intra-class invariance by dynamically mining meaningful pseudo positive samples from both mini-batches and a memory bank during training. Extensive experiments demonstrate that InsCLR achieves similar or even better performance than the state-of-the-art SSL methods on instance retrieval. Code is available at https://github.com/zeludeng/insclr.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

Novelty detection aims to automatically identify out-of-distribution (OOD) data, without any prior knowledge of them. It is a critical step in data monitoring, behavior analysis and other applications, helping enable continual learning in the field. Conventional methods of OOD detection perform multi-variate analysis on an ensemble of data or features, and usually resort to the supervision with OOD data to improve the accuracy. In reality, such supervision is impractical as one cannot anticipate the anomalous data. In this paper, we propose a novel, self-supervised approach that does not rely on any pre-defined OOD data: (1) The new method evaluates the Mahalanobis distance of the gradients between the in-distribution and OOD data. (2) It is assisted by a self-supervised binary classifier to guide the label selection to generate the gradients, and maximize the Mahalanobis distance. In the evaluation with multiple datasets, such as CIFAR-10, CIFAR-100, SVHN and TinyImageNet, the proposed approach consistently outperforms state-of-the-art supervised and unsupervised methods in the area under the receiver operating characteristic (AUROC) and area under the precision-recall curve (AUPR) metrics. We further demonstrate that this detector is able to accurately learn one OOD class in continual learning.

poster session 1 @ blue 5, poster session 11 @ blue 5, poster session 1, poster session 11

Generalized zero-shot learning (GZSL) is a learning technique to identify unseen classes by exploiting attributes. In this paper, we propose a new GZSL technique exploiting attribute-related image features. While conventional GZSL approaches use image features containing attribute-irrelevant information for the classification, our approach, henceforth referred to as semantic feature extraction-based GZSL (SE-GZSL), exploits semantic features, features containing only attribute-related information, to remove the interference caused by attribute-irrelevant information. To extract semantic features, we introduce 1) mutual information (MI)-based loss to capture all the attribute-related information in the image features and 2) similarity-based loss to remove attribute-irrelevant information. From extensive experiments on various benchmark datasets, we show that SE-GZSL outperforms conventional GZSL techniques.

poster session 1 @ blue 4, poster session 11 @ blue 4, poster session 1, poster session 11

We posit a new mechanism for cooperation in multi-agent reinforcement learning (MARL) based upon any nonlinear function of the team's long-term state-action occupancy measure, i.e., a \emph{general utility}. This subsumes the cumulative return but also allows one to incorporate risk-sensitivity, exploration, and priors. % We derive the {\bf D}ecentralized {\bf S}hadow Reward {\bf A}ctor-{\bf C}ritic (DSAC) in which agents alternate between policy evaluation (critic), weighted averaging with neighbors (information mixing), and local gradient updates for their policy parameters (actor). DSAC augments the classic critic step by requiring agents to (i) estimate their local occupancy measure in order to (ii) estimate the derivative of the local utility with respect to their occupancy measure, i.e., the ``shadow reward". DSAC converges to $\epsilon$-stationarity in $\mathcal{O}(1/\epsilon^{2.5})$ (Theorem \ref{theorem:final}) or faster $\mathcal{O}(1/\epsilon^{2})$ (Corollary \ref{corollary:communication}) steps with high probability, depending on the amount of communications. We further establish the non-existence of spurious stationary points for this problem, that is, DSAC finds the globally optimal policy (Corollary \ref{corollary:global}). Experiments demonstrate the merits of goals beyond the cumulative return in cooperative MARL.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

Despite the great progress in video understanding made by deep convolutional neural networks, feature representation learned by existing methods may be biased to static visual cues. To address this issue, we propose a novel method to suppress static visual cues (SSVC) based on probabilistic analysis for self-supervised video representation learning. In our method, video frames are first encoded to obtain latent variables under standard normal distribution via normalizing flows. By modelling static factors in a video as a random variable, the conditional distribution of each latent variable becomes shifted and scaled normal. Then, the less-varying latent variables along time are selected as static cues and suppressed to generate motion-preserved videos. Finally, positive pairs are constructed by motion-preserved videos for contrastive learning to alleviate the problem of representation bias to static cues. The less-biased video representation can be better generalized to various downstream tasks. Extensive experiments on publicly available benchmarks demonstrate that the proposed method outperforms the state of the art when only single RGB modality is used for pre-training.

poster session 1 @ red 1, poster session 11 @ red 1, oral session 1 @ red 1, poster session 1, poster session 11, oral session 1

We study regret minimization in finite horizon tabular Markov decision processes (MDPs) under the constraints of differential privacy (DP). This is motivated by the widespread applications of reinforcement learning (RL) in real-world sequential decision making problems, where protecting users' sensitive and private information is becoming paramount. We consider two variants of DP -- joint DP (JDP), where a centralized agent is responsible for protecting users' sensitive data and local DP (LDP), where information needs to be protected directly on the user side. We first propose two general frameworks -- one for policy optimization and another for value iteration -- for designing private, optimistic RL algorithms. We then instantiate these frameworks with suitable privacy mechanisms to satisfy JDP and LDP requirements, and simultaneously obtain sublinear regret guarantees. The regret bounds show that under JDP, the cost of privacy is only a lower order additive term, while for a stronger privacy protection under LDP, the cost suffered is multiplicative. Finally, the regret bounds are obtained by a unified analysis, which, we believe, can be extended beyond tabular MDPs.

poster session 1 @ blue 1, poster session 8 @ blue 1, oral session 1 @ blue 1, poster session 1, poster session 8, oral session 1

We empirically investigate the effect of class manifold entanglement and the intrinsic and extrinsic dimensionality of the data distribution on the sample complexity of supervised classification with deep ReLU networks. We separate the effect of entanglement and intrinsic dimensionality and show statistically for artificial and real-world image datasets that the intrinsic dimensionality and the entanglement have an interdependent effect on the sample complexity. Low levels of entanglement lead to low increases of the sample complexity when the intrinsic dimensionality is increased, while for high levels of entanglement the impact of the intrinsic dimensionality increases as well. Further, we show that in general the sample complexity is primarily due to the entanglement and only secondarily due to the intrinsic dimensionality of the data distribution.

poster session 1 @ blue 3, poster session 12 @ blue 3, poster session 1, poster session 12

3D delineation of anatomical structures is a cardinal goal in medical imaging analysis. Prior to deep learning, statistical shape models that imposed anatomical constraints and produced high quality surfaces were a core technology. Prior to deep learning, statistical shape models that imposed anatomical constraints and produced high quality surfaces were a core technology. Today fully-convolutional networks (FCNs), while dominant, do not offer these capabilities. We present deep implicit statistical shape models (DISSMs), a new approach to delineation that marries the representation power of convolutional neural networks (CNNs) with the robustness of SSMs. DISSMs use a deep implicit surface representation to produce a compact and descriptive shape latent space that permits statistical models of anatomical variance. To reliably fit anatomically plausible shapes to an image, we introduce a novel rigid and non-rigid pose estimation pipeline that is modelled as a Markov decision process(MDP). We outline a training regime that includes inverted episodic training and a deep realization of marginal space learning (MSL). Intra-dataset experiments on the task of pathological liver segmentation demonstrate that DISSMs can perform more robustly than three leading FCN models, including nnU-Net: reducing the mean Hausdorff distance (HD) by 7.7-14.3mm and improving the worst case Dice-Sorensen coefficient (DSC) by 1.2-2.3%. More critically, cross-dataset experiments on a dataset directly reflecting clinical deployment scenarios demonstrate that DISSMs improve the mean DSC and HD by 3.5-5.9% and 12.3-24.5mm, respectively, and the worst-case DSC by 5.4-7.3%. These improvements are over and above any benefits from representing delineations with high-quality surface.

poster session 1 @ red 2, poster session 10 @ red 2, oral session 10 @ red 2, poster session 1, poster session 10, oral session 10

Online learning algorithms have become a ubiquitous tool in the machine learning toolbox and are frequently used in small, resource-constraint environments. Among the most successful online learning methods are Decision Tree (DT) ensembles. DT ensembles provide excellent performance while adapting to changes in the data, but they are not resource efficient. Incremental tree learners keep adding new nodes to the tree but never remove old ones increasing the memory consumption over time. Gradient-based tree learning, on the other hand, requires the computation of gradients over the entire tree which is costly for even moderately sized trees. In this paper, we propose a novel memory-efficient online classification ensemble called shrub ensembles for resource-constraint systems. Our algorithm trains small to medium-sized decision trees on small windows and uses stochastic proximal gradient descent to learn the ensemble weights of these `shrubs'. We provide a theoretical analysis of our algorithm and include an extensive discussion on the behavior of our approach in the online setting. In a series of 2~959 experiments on 12 different datasets, we compare our method against 8 state-of-the-art methods. Our Shrub Ensembles retain an excellent performance even when only little memory is available. We show that SE offers a better accuracy-memory trade-off in 7 of 12 cases, while having a statistically significant better performance than most other methods. Our implementation is available under https://github.com/sbuschjaeger/se-online .

poster session 1 @ blue 3, poster session 12 @ blue 3, oral session 1 @ blue 3, poster session 1, poster session 12, oral session 1

We consider a setting where a large number of agents are all interested in attending some public resource of limited capacity. Attendance is thus allotted by lottery. If agents arrive individually, then randomly choosing the agents – one by one - is a natural, fair and efficient solution. We consider the case where agents are organized in groups (e.g. families, friends), the members of each of which must all be admitted together. We study the question of how best to design such lotteries. We first establish the desired properties of such lotteries, in terms of fairness and efficiency, and define the appropriate notions of strategy proofness (providing that agents cannot gain by misrepresenting the true groups, e.g. joining or splitting groups). We establish inter-relationships between the different properties, proving properties that cannot be fulfilled simultaneously (e.g. leximin optimality and strong group stratagy proofness). Our main contribution is a polynomial mechanism for the problem, which guarantees many of the desired properties, including: leximin optimality, Pareto-optimality, anonymity, group strategy proofness, and adjunctive strategy proofness (which provides that no benefit can be obtained by registering additional - uninterested or bogus - individuals). The mechanism approximates the utilitarian optimum to within a factor of 2, which, we prove, is optimal for any mechanism that guarantees any one of the following properties: egalitarian welfare optimality, leximin optimality, envyfreeness, and adjunctive strategy proofness.

poster session 1 @ blue 6, poster session 12 @ blue 6, poster session 1, poster session 12

This work introduces Fractional Activation Linear Units (FALUs), a flexible generalization of adaptive activation functions. Leveraging principles from fractional calculus, FALUs define a diverse family of activation functions that encompass many traditional and state-of-the-art activation functions, including the Sigmoid, ReLU, Swish, GELU, and Gaussian functions, as well as a large variety of smooth interpolations between these functions. Our technique introduces only a small number of additional trainable parameters, and requires no additional specialized optimization or initialization procedures. For this reason, FALUs present a seamless and rich automated solution to the problem of activation function optimization. Through experiments on a variety of conventional tasks and network architectures, we demonstrate the effectiveness of FALUs when compared to traditional and state-of-the-art activation functions.

poster session 1 @ blue 2, poster session 11 @ blue 2, poster session 1, poster session 11

Model-free reinforcement learning (RL) algorithms are known to be memory and computation-efficient compared with model-based approaches. This paper presents a model-free RL algorithm for infinite-horizon average-reward Constrained Markov Decision Processes (CMDPs), which achieves sub-linear regret and zero constraint violation. To the best of our knowledge, this is the first {\em model-free} algorithm for general CMDPs in the infinite-horizon average-reward setting with provable guarantees.

poster session 1 @ blue 1, poster session 8 @ blue 1, poster session 1, poster session 8

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

In recent years, pretrained language models have revolutionized the NLP world, while achieving state of the art performance in various downstream tasks. However, in many cases, these models do not perform well when labeled data is scarce and the model is expected to perform in the zero or few shot setting. Recently, several works have shown that continual pretraining or performing a second phase of pretraining (inter-training) which is better aligned with the downstream task, can lead to improved results, especially in the scarce data setting. Here, we propose to leverage sentiment-carrying discourse markers to generate large-scale weakly-labeled data, which in turn can be used to adapt language models for sentiment analysis. Extensive experimental results show the value of our approach on various benchmark datasets, including the finance domain. Code, models and data are available at https://github.com/ibm/tslm-discourse-markers.

poster session 1 @ red 4, poster session 8 @ red 4, poster session 1, poster session 8

Large capacity deep learning models are often prone to a high generalization gap when trained with a limited amount of labeled training data. A recent class of methods to address this problem uses various ways to construct a new training sample by mixing a pair (or more) of training samples. We propose PatchUp, a hidden state block-level regularization technique for Convolutional Neural Networks (CNNs), that is applied on selected contiguous blocks of feature maps from a random pair of samples. Our approach improves the robustness of CNN models against the manifold intrusion problem that may occur in other state-of-the-art mixing approaches. Moreover, since we are mixing the contiguous block of features in the hidden space, which has more dimensions than the input space, we obtain more diverse samples for training towards different dimensions. Our experiments on CIFAR10/100, SVHN, Tiny-ImageNet, and ImageNet using ResNet architectures including PreActResnet18/34, WRN-28-10, ResNet101/152 models show that PatchUp improves upon, or equals, the performance of current state-of-the-art regularizers for CNNs. We also show that PatchUp can provide a better generalization to deformed samples and is more robust against adversarial attacks.

poster session 1 @ red 3, poster session 11 @ red 3, poster session 1, poster session 11