IJCAI 2021 论文列表

We introduce "argument-incomplete Abstract Argumentation Frameworks with dependencies", that extend the traditional abstract argumentation reasoning to the case where some arguments are uncertain and correlated through logical dependencies (such as mutual exclusion, implication, etc.). We characterize the complexities of the problems DSAT of deciding the satisfiability of the dependencies and PDVER of verifying extensions, and show how they depend on the forms of dependencies and, for PDVER, also on the semantics of the extensions.

In a recent paper Belle and Lakemeyer proposed the logic DS, a probabilistic extension of a modal variant of the situation calculus with a model of belief based on weighted possible worlds. Among other things, they were able to precisely capture the beliefs of a probabilistic knowledge base in terms of the concept of only-believing. While intuitively appealing, the logic has a number of shortcomings. Perhaps the most severe is the limited expressiveness in that degrees of belief are restricted to constant rational numbers, which makes it impossible to express arbitrary belief distributions. In this paper we will address this and other shortcomings by extending the language and modifying the semantics of belief and only-believing. Among other things, we will show that belief retains many but not all of the properties of DS. Moreover, it turns out that only-believing arbitrary sentences, including those mentioning belief, is uniquely satisfiable in our logic. For an interesting class of knowledge bases we also show how reasoning about beliefs and meta-beliefs after performing noisy actions and sensing can be reduced to reasoning about the initial beliefs of an agent using a form of regression.

We study the semantics of knowledge in strategic reasoning. Most existing works either implicitly assume that agents do not know one another’s strategies, or that all strategies are known to all; and some works present inconsistent mixes of both features. We put forward a novel semantics for Strategy Logic with Knowledge that cleanly models whose strategies each agent knows. We study how adopting this semantics impacts agents’ knowledge and strategic ability, as well as the complexity of the model-checking problem.

When demand increases beyond the system capacity, riders in ride-hailing/ride-sharing systems often experience long waiting time, resulting in poor customer satisfaction. This paper proposes a spatio-temporal pricing framework (AP-RTRS) to alleviate this challenge and shows how it naturally complements state-of-the-art dispatching and routing algorithms. Specifically, the pricing optimization model regulates demand to ensure that every rider opting to use the system is served within reason-able time: it does so either by reducing demand to meet the capacity constraints or by prompting potential riders to postpone service to a later time. The pricing model is a model-predictive control algorithm that works at a coarser temporal and spatial granularity compared to the real-time dispatching and routing, and naturally integrates vehicle relocations. Simulation experiments indicate that the pricing optimization model achieves short waiting times without sacrificing revenues and geographical fairness.

Extension-based semantics in abstract argumentation provide a criterion to determine whether a set of arguments is acceptable or not. In this paper, we present the notion of extension-ranking semantics, which determines a preordering over sets of arguments, where one set is deemed more plausible than another if it is somehow more acceptable. We obtain extension-based semantics as a special case of this new approach, but it also allows us to make more fine-grained distinctions, such as one set being "more complete'' or "more admissible'' than another. We define a number of general principles to classify extension-ranking semantics and develop concrete approaches. We also study the relation between extension-ranking semantics and argument-ranking based semantics, which rank individual arguments instead of sets of arguments.

Human-Object Interaction (HOI) detection devotes to learn how humans interact with surrounding objects. Latest end-to-end HOI detectors are short of relation reasoning, which leads to inability to learn HOI-specific interactive semantics for predictions. In this paper, we therefore propose novel relation reasoning for HOI detection. We first present a progressive Relation-aware Frame, which brings a new structure and parameter sharing pattern for interaction inference. Upon the frame, an Interaction Intensifier Module and a Correlation Parsing Module are carefully designed, where: a) interactive semantics from humans can be exploited and passed to objects to intensify interactions, b) interactive correlations among humans, objects and interactions are integrated to promote predictions. Based on modules above, we construct an end-to-end trainable framework named Relation Reasoning Network (abbr. RR-Net). Extensive experiments show that our proposed RR-Net sets a new state-of-the-art on both V-COCO and HICO-DET benchmarks and improves the baseline about 5.5% and 9.8% relatively, validating that this first effort in exploring relation reasoning and integrating interactive semantics has brought obvious improvement for end-to-end HOI detection.

Unsupervised anomaly detection plays a crucial role in many critical applications. Driven by the success of deep learning, recent years have witnessed growing interests in applying deep neural networks (DNNs) to anomaly detection problems. A common approach is using autoencoders to learn a feature representation for the normal observations in the data. The reconstruction error of the autoencoder is then used as outlier scores to detect the anomalies. However, due to the high complexity brought upon by the over-parameterization of DNNs, the reconstruction error of the anomalies could also be small, which hampers the effectiveness of these methods. To alleviate this problem, we propose a robust framework using collaborative autoencoders to jointly identify normal observations from the data while learning its feature representation. We investigate the theoretical properties of the framework and empirically show its outstanding performance as compared to other DNN-based methods. Our experimental results also show the resiliency of the framework to missing values compared to other baseline methods.

We introduce Recurrent All-Pairs Field Transforms (RAFT), a new deep network architecture for optical flow. RAFT extracts per-pixel features, builds multi-scale 4D correlation volumes for all pairs of pixels, and iteratively updates a flow field through a recurrent unit that performs lookups on the correlation volumes. RAFT achieves state-of-the-art performance on the KITTI and Sintel datasets. In addition, RAFT has strong cross-dataset generalization as well as high efficiency in inference time, training speed, and parameter count.

Qualitative Spatial & Temporal Reasoning (QSTR) is a major field of study in Symbolic AI that deals with the representation and reasoning of spatio- temporal information in an abstract, human-like manner. We survey the current status of QSTR from a viewpoint of reasoning approaches, and identify certain future challenges that we think that, once overcome, will allow the field to meet the demands of and adapt to real-world, dynamic, and time-critical applications of highly active areas such as machine learning and data mining.

In their AAMAS 2020 paper, Szufa et al. presented a "map of elections" that visualizes a set of 800 elections generated from various statistical cultures. While similar elections are grouped together on this map, there is no obvious interpretation of the elections' positions. We provide such an interpretation by introducing four canonical “extreme” elections, acting as a compass on the map. We use them to analyze both a dataset provided by Szufa et al. and a number of real-life elections. In effect, we find a new parameterization of the Mallows model, based on measuring the expected swap distance from the central preference order, and show that it is useful for capturing real-life scenarios.

Spiking Neural Networks (SNNs) have been attached great importance due to their biological plausibility and high energy-efficiency on neuromorphic chips. As these chips are usually resource-constrained, the compression of SNNs is thus crucial along the road of practical use of SNNs. Most existing methods directly apply pruning approaches in artificial neural networks (ANNs) to SNNs, which ignore the difference between ANNs and SNNs, thus limiting the performance of the pruned SNNs. Besides, these methods are only suitable for shallow SNNs. In this paper, inspired by synaptogenesis and synapse elimination in the neural system, we propose gradient rewiring (Grad R), a joint learning algorithm of connectivity and weight for SNNs, that enables us to seamlessly optimize network structure without retraining. Our key innovation is to redefine the gradient to a new synaptic parameter, allowing better exploration of network structures by taking full advantage of the competition between pruning and regrowth of connections. The experimental results show that the proposed method achieves minimal loss of SNNs' performance on MNIST and CIFAR-10 datasets so far. Moreover, it reaches a ~3.5% accuracy loss under unprecedented 0.73% connectivity, which reveals remarkable structure refining capability in SNNs. Our work suggests that there exists extremely high redundancy in deep SNNs. Our codes are available at https://github.com/Yanqi-Chen/Gradient-Rewiring.

Robustness of decision rules to shifts in the data-generating process is crucial to the successful deployment of decision-making systems. Such shifts can be viewed as interventions on a causal graph, which capture (possibly hypothetical) changes in the data-generating process, whether due to natural reasons or by the action of an adversary. We consider causal Bayesian networks and formally define the interventional robustness problem, a novel model-based notion of robustness for decision functions that measures worst-case performance with respect to a set of interventions that denote changes to parameters and/or causal influences. By relying on a tractable representation of Bayesian networks as arithmetic circuits, we provide efficient algorithms for computing guaranteed upper and lower bounds on the interventional robustness probabilities. Experimental results demonstrate that the methods yield useful and interpretable bounds for a range of practical networks, paving the way towards provably causally robust decision-making systems.

Referring Expression Comprehension (REC) has become one of the most important tasks in visual reasoning, since it is an essential step for many vision-and-language tasks such as visual question answering. However, it has not been widely used in many downstream tasks because it suffers 1) two-stage methods exist heavy computation cost and inevitable error accumulation, and 2) one-stage methods have to depend on lots of hyper-parameters (such as anchors) to generate bounding box. In this paper, we present a proposal-free one-stage (PFOS) model that is able to regress the region-of-interest from the image, based on a textual query, in an end-to-end manner. Instead of using the dominant anchor proposal fashion, we directly take the dense-grid of image as input for a cross-attention transformer that learns grid-word correspondences. The final bounding box is predicted directly from the image without the time-consuming anchor selection process that previous methods suffer. Our model achieves the state-of-the-art performance on four referring expression datasets with higher efficiency, comparing to previous best one-stage and two-stage methods.

It is desirable to include more controllable attributes to enhance the diversity of generated responses in open-domain dialogue systems. However, existing methods can generate responses with only one controllable attribute or lack a flexible way to generate them with multiple controllable attributes. In this paper, we propose a Progressively trained Hierarchical Encoder-Decoder (PHED) to tackle this task. More specifically, PHED deploys Conditional Variational AutoEncoder (CVAE) on Transformer to include one aspect of attributes at one stage. A vital characteristic of the CVAE is to separate the latent variables at each stage into two types: a global variable capturing the common semantic features and a specific variable absorbing the attribute information at that stage. PHED then couples the CVAE latent variables with the Transformer encoder and is trained by minimizing a newly derived ELBO and controlled losses to produce the next stage's input and produce responses as required. Finally, we conduct extensive evaluations to show that PHED significantly outperforms the state-of-the-art neural generation models and produces more diverse responses as expected.

Given a specification φ(X, Y ) over inputs X and output Y and defined over a background theory T, the problem of program synthesis is to design a program f such that Y = f (X), satisfies the specification φ. Over the past decade, syntax-guided synthesis (SyGuS) has emerged as a dominant approach to program synthesis where in addition to the specification φ, the end-user also specifies a grammar L to aid the underlying synthesis engine. This paper investigates the feasibility of synthesis techniques without grammar, a sub-class defined as T constrained synthesis. We show that T-constrained synthesis can be reduced to DQF(T),i.e., to the problem of finding a witness of a dependency quantified formula modulo theory. When the underlying theory is the theory of bitvectors, the corresponding DQF problem can be further reduced to Dependency Quantified Boolean Formulas (DQBF). We rely on the progress in DQBF solving to design DQBF-based synthesizers that outperform the domain-specific program synthesis techniques; thereby positioning DQBF as a core representation language for program synthesis. Our empirical analysis shows that T-constrained synthesis can achieve significantly better performance than syntax-guided approaches. Furthermore, the general-purpose DQBF solvers perform on par with domain-specific synthesis techniques.

Understanding the behavior of learned classifiers is an important task, and various black-box explanations, logical reasoning approaches, and model-specific methods have been proposed. In this paper, we introduce probabilistic sufficient explanations, which formulate explaining an instance of classification as choosing the "simplest" subset of features such that only observing those features is "sufficient" to explain the classification. That is, sufficient to give us strong probabilistic guarantees that the model will behave similarly when all features are observed under the data distribution. In addition, we leverage tractable probabilistic reasoning tools such as probabilistic circuits and expected predictions to design a scalable algorithm for finding the desired explanations while keeping the guarantees intact. Our experiments demonstrate the effectiveness of our algorithm in finding sufficient explanations, and showcase its advantages compared to Anchors and logical explanations.

We study the differentially private (DP) stochastic nonconvex optimization with a focus on its under-studied utility measures in terms of the expected excess empirical and population risks. While the excess risks are extensively studied for convex optimization, they are rarely studied for nonconvex optimization, especially the expected population risk. For the convex case, recent studies show that it is possible for private optimization to achieve the same order of excess population risk as to the nonprivate optimization under certain conditions. It still remains an open question for the nonconvex case whether such ideal excess population risk is achievable. In this paper, we progress towards an affirmative answer to this open problem: DP nonconvex optimization is indeed capable of achieving the same excess population risk as to the nonprivate algorithm in most common parameter regimes, under certain conditions (i.e., well-conditioned nonconvexity). We achieve such improved utility rates compared to existing results by designing and analyzing the stagewise DP-SGD with early momentum algorithm. We obtain both excess empirical risk and excess population risk to achieve differential privacy. Our algorithm also features the first known results of excess and population risks for DP-SGD with momentum. Experiment results on both shallow and deep neural networks when respectively applied to simple and complex real datasets corroborate the theoretical results.

Text generation has become one of the most important yet challenging tasks in natural language processing (NLP). The resurgence of deep learning has greatly advanced this field by neural generation models, especially the paradigm of pretrained language models (PLMs). In this paper, we present an overview of the major advances achieved in the topic of PLMs for text generation. As the preliminaries, we present the general task definition and briefly describe the mainstream architectures of PLMs for text generation. As the core content, we discuss how to adapt existing PLMs to model different input data and satisfy special properties in the generated text. We further summarize several important fine-tuning strategies for text generation. Finally, we present several future directions and conclude this paper. Our survey aims to provide text generation researchers a synthesis and pointer to related research.

In recommender systems, the cold-start problem is a critical issue. To alleviate this problem, an emerging direction adopts meta-learning frameworks and achieves success. Most existing works aim to learn globally shared prior knowledge across all users so that it can be quickly adapted to a new user with sparse interactions. However, globally shared prior knowledge may be inadequate to discern users’ complicated behaviors and causes poor generalization. Therefore, we argue that prior knowledge should be locally shared by users with similar preferences who can be recognized by social relations. To this end, in this paper, we propose a Preference-Adaptive Meta-Learning approach (PAML) to improve existing meta-learning frameworks with better generalization capacity. Specifically, to address two challenges imposed by social relations, we first identify reliable implicit friends to strengthen a user’s social relations based on our defined palindrome paths. Then, a coarse-fine preference modeling method is proposed to leverage social relations and capture the preference. Afterwards, a novel preference-specific adapter is designed to adapt the globally shared prior knowledge to the preference-specific knowledge so that users who have similar tastes share similar knowledge. We conduct extensive experiments on two publicly available datasets. Experimental results validate the power of social relations and the effectiveness of PAML.