IJCAI 2023 论文列表

Recent graph-based methods achieve significant success in multivariate time series modeling and forecasting due to their ability to handle relationships among time series variables. However, only pairwise relationships are considered in most existing works. They ignore beyond-pairwise relationships and their potential categories in practical scenarios, which leads to incomprehensive relationship learning for multivariate time series forecasting. In this paper, we present ReMo, a Relational Modeling-based method, to promote fine-grained relational learning among multivariate time series data. Firstly, by treating time series variables and complex relationships as nodes and hyperedges, we extract multi-view hypergraphs from data to capture beyond-pairwise relationships. Secondly, a novel hypergraph message passing strategy is designed to characterize both nodes and hyperedges by inferring the potential categories of relationships and further distinguishing their impacts on time series variables. By integrating these two modules into the time series forecasting framework, ReMo effectively improves the performance of multivariate time series forecasting. The experimental results on seven commonly used datasets from different domains demonstrate the superiority of our model.

The power of norms in both human societies and sociotechnical systems arises from the facts that (1) societal norms, including laws and policies, characterize acceptable behavior in high-level terms and (2) they are not hard controls and can be deviated from. Thus, the design of responsibly autonomous agents faces an essential tension: these agents must both (1) respect applicable norms and (2) deviate from those norms when blindly following them may lead to diminished outcomes. We propose a conceptual foundation for norm deviation. As a guiding framework, we adopt Habermas's theory of communicative action comprising objective, subjective, and practical validity claims regarding the suitability of deviation. Our analysis thus goes beyond previous studies of norm deviation and yields reasoning guidelines uniting norms and values by which to develop responsible agents.

Recent work in algorithmic mechanism design focuses on designing mechanisms for agents with bounded rationality, modifying the constraints that must be satisfied in order to achieve incentive compatibility. Starting with Li's strengthening of strategyproofness, obvious strategyproofness (OSP) requires truthtelling to be "obvious" over dishonesty, roughly meaning that the worst outcome from truthful actions must be no worse than the best outcome for dishonest ones. A celebrated result for dominant-strategy incentive-compatible mechanisms that allows us to restrict attention to direct mechanisms, known as the revelation principle, does not hold for OSP: the implementation details matter for the obvious incentive properties of the mechanism. Studying agent strategies in real-life mechanisms, Troyan and Morrill introduce a relaxation of strategyproofness known as non-obvious manipulability, which only requires comparing certain extrema of the agents' utility functions in order for a mechanism to be incentive-compatible. Specifically a mechanism is not obviously manipulable (NOM) if the best and worst outcomes when acting truthfully are no worse than the best and worst outcomes when acting dishonestly. In this work we first extend the cycle monotonicity framework for direct-revelation NOM mechanism design to indirect mechanisms. We then apply this to two settings, single-parameter agents and mechanisms for two agents in which one has a two-value domain, and show that under these models the revelation principle holds: direct mechanisms are just as powerful as indirect ones.

Multispectral photometric stereo (MPS) aims at recovering the surface normal of a scene from a single-shot multispectral image captured under multispectral illuminations. Existing MPS methods adopt the Lambertian reflectance model to make the problem tractable, but it greatly limits their application to real-world surfaces. In this paper, we propose a deep neural network named NeuralMPS to solve the MPS problem under non-Lambertian spectral reflectances. Specifically, we present a spectral reflectance decomposition model to disentangle the spectral reflectance into a geometric component and a spectral component. With this decomposition, we show that the MPS problem for surfaces with a uniform material is equivalent to the conventional photometric stereo (CPS) with unknown light intensities. In this way, NeuralMPS reduces the difficulty of the non-Lambertian MPS problem by leveraging the well-studied non-Lambertian CPS methods. Experiments on both synthetic and real-world scenes demonstrate the effectiveness of our method.

For the fundamental problem of fairly dividing a set of indivisible items among agents, envy-freeness up to any item (EFX) and maximin fairness (MMS) are arguably the most compelling fairness concepts proposed till now. Unfortunately, despite significant efforts over the past few years, whether EFX allocations always exist is still an enigmatic open problem, let alone their efficient computation. Furthermore, today we know that MMS allocations are not always guaranteed to exist. These facts weaken the usefulness of both EFX and MMS, albeit their appealing conceptual characteristics. We propose two alternative fairness concepts—called epistemic EFX (EEFX) and minimum EFX value fairness (MXS)---inspired by EFX and MMS. For both, we explore their relationships to well-studied fairness notions and, more importantly, prove that EEFX and MXS allocations always exist and can be computed efficiently for additive valuations. Our results justify that the new fairness concepts are excellent alternatives to EFX and MMS.

This paper addresses the weighted vertex coloring problem (WVCP) which is an NP-hard variant of the graph coloring problem with various applications. Given a vertex-weighted graph, the problem consists of partitioning vertices in independent sets (colors) so as to minimize the sum of the maximum weights of the colors. We first present an iterative procedure to reduce the size of WVCP instances and prove new upper bounds on the objective value and the number of colors. Alternative constraint programming models are then introduced which rely on primal and dual encodings of the problem and use symmetry breaking constraints. A large number of experiments are conducted on benchmark instances. We analyze the impact of using specific bounds to reduce the search space and speed up the exact resolution of instances. New optimality proofs are reported for some benchmark instances.

We study the problem of fairly and efficiently allocating indivisible chores among agents with additive disutility functions. We consider the widely used envy-based fairness properties of EF1 and EFX in conjunction with the efficiency property of fractional Pareto-optimality (fPO). Existence (and computation) of an allocation that is simultaneously EF1/EFX and fPO are challenging open problems, and we make progress on both of them. We show the existence of an allocation that is - EF1 + fPO, when there are three agents, - EF1 + fPO, when there are at most two disutility functions, - EFX + fPO, for three agents with bivalued disutility functions. These results are constructive, based on strongly polynomial-time algorithms. We also investigate non-existence and show that an allocation that is EFX+fPO need not exist, even for two agents.

One of the ultimate goals of Artificial Intelligence is to assist humans in complex decision making. A promising direction for achieving this goal is Neuro-Symbolic AI, which aims to combine the interpretability of symbolic techniques with the ability of deep learning to learn from raw data. However, most current approaches require manually engineered symbolic knowledge, and where end-to-end training is considered, such approaches are either restricted to learning definite programs, or are restricted to training binary neural networks. In this paper, we introduce Neuro-Symbolic Inductive Learner (NSIL), an approach that trains a general neural network to extract latent concepts from raw data, whilst learning symbolic knowledge that maps latent concepts to target labels. The novelty of our approach is a method for biasing the learning of symbolic knowledge, based on the in-training performance of both neural and symbolic components. We evaluate NSIL on three problem domains of different complexity, including an NP-complete problem. Our results demonstrate that NSIL learns expressive knowledge, solves computationally complex problems, and achieves state-of-the-art performance in terms of accuracy and data efficiency. Code and technical appendix: https://github.com/DanCunnington/NSIL

Models computed using deep learning have been effectively applied to tackle various problems in many disciplines. Yet, the predictions of these models are often at most post-hoc and locally explainable. In contrast, class expressions in description logics are ante-hoc and globally explainable. Although state-of-the-art symbolic machine learning approaches are being successfully applied to learn class expressions, their application at large scale has been hindered by their impractical runtimes. Arguably, the reliance on myopic heuristic functions contributes to this limitation. We propose a novel neuro-symbolic class expression learning model, DRILL, to mitigate this limitation. By learning non-myopic heuristic functions with deep Q-learning, DRILL efficiently steers the standard search procedure in a quasi-ordered search space towards goal states. Our extensive experiments on 4 benchmark datasets and 390 learning problems suggest that DRILL converges to goal states at least 2.7 times faster than state-of-the-art models on all learning problems. The results of our statistical significance test confirms that DRILL converges to goal states significantly faster (p-value <1%) than state-of-the-art models on all benchmark datasets. We provide an open-source implementation of DRILL, including pre-trained models, training and evaluation scripts.

Recent work on deep clustering has found new promising methods also for constrained clustering problems. Their typically pairwise constraints often can be used to guide the partitioning of the data. Many problems however, feature cluster-level constraints, e.g. the Capacitated Clustering Problem (CCP), where each point has a weight and the total weight sum of all points in each cluster is bounded by a prescribed capacity. In this paper we propose a new method for the CCP, Neural Capacited Clustering, that learns a neural network to predict the assignment probabilities of points to cluster centers from a data set of optimal or near optimal past solutions of other problem instances. During inference, the resulting scores are then used in an iterative k-means like procedure to refine the assignment under capacity constraints. In our experiments on artificial data and two real world datasets our approach outperforms several state-of-the-art mathematical and heuristic solvers from the literature. Moreover, we apply our method in the context of a cluster-first-route-second approach to the Capacitated Vehicle Routing Problem (CVRP) and show competitive results on the well-known Uchoa benchmark.

In this paper, we introduce Neural Probabilistic Soft Logic (NeuPSL), a novel neuro-symbolic (NeSy) framework that unites state-of-the-art symbolic reasoning with the low-level perception of deep neural networks. To model the boundary between neural and symbolic representations, we propose a family of energy-based models, NeSy Energy-Based Models, and show that they are general enough to include NeuPSL and many other NeSy approaches. Using this framework, we show how to seamlessly integrate neural and symbolic parameter learning and inference in NeuPSL. Through an extensive empirical evaluation, we demonstrate the benefits of using NeSy methods, achieving upwards of 30% improvement over independent neural network models. On a well-established NeSy task, MNIST-Addition, NeuPSL demonstrates its joint reasoning capabilities by outperforming existing NeSy approaches by up to 10% in low-data settings. Furthermore, NeuPSL achieves a 5% boost in performance over state-of-the-art NeSy methods in a canonical citation network task with up to a 40 times speed up.

Sequence labeling serves as the most commonly used scheme for Chinese named entity recognition(NER). However, traditional sequence labeling methods classify tokens within an entity into different classes according to their positions. As a result, different tokens in the same entity may be learned with representations that are isolated and unrelated in target representation space, which could finally negatively affect the subsequent performance of token classification. In this paper, we point out and define this problem as Entity Representation Segmentation in Label-semantics. And then we present NerCo: Named entity recognition with Contrastive learning, a novel NER framework which can better exploit labeled data and avoid the above problem. Following the pretrain-finetune paradigm, NerCo firstly guides the encoder to learn powerful label-semantics based representations by gathering the encoded token representations of the same Semantic Class while pushing apart that of different. Subsequently, NerCo finetunes the learned encoder for final entity prediction. Extensive experiments on several datasets demonstrate that our framework can consistently improve the baseline and achieve state-of-the-art performance.

Markov Logic Networks (MLN) are used for reasoning on uncertain and inconsistent temporal data. We proposed the TMLN (Temporal Markov Logic Network) which extends them with sorts/types, weights on rules and facts, and various temporal consistencies. The NeoMaPy framework integrates it as a knowledge graph based on conflict graphs which offers flexibility for reasoning with parametric Maximum A Posteriori (MAP) inferences, efficiency with an optimistic heuristic and interactive graph visualization for results explanation.

There are increasing demands for understanding deep neural networks' (DNNs) behavior spurred by growing security and/or transparency concerns. Due to multi-layer nonlinearity of the deep neural network architectures, explaining DNN predictions still remains as an open problem, preventing us from gaining a deeper understanding of the mechanisms. To enhance the explainability of DNNs, we estimate the input feature's attributions to the prediction task using divergence and flux. Inspired by the divergence theorem in vector analysis, we develop a novel Negative Flux Aggregation (NeFLAG) formulation and an efficient approximation algorithm to estimate attribution map. Unlike the previous techniques, ours doesn't rely on fitting a surrogate model nor need any path integration of gradients. Both qualitative and quantitative experiments demonstrate a superior performance of NeFLAG in generating more faithful attribution maps than the competing methods. Our code is available at https://github.com/xinli0928/NeFLAG.

Developing digital sound synthesizers is crucial to the music industry as it provides a low-cost way to produce high-quality sounds with rich timbres. Existing traditional synthesizers often require substantial expertise to determine the overall framework of a synthesizer and the parameters of submodules. Since expert knowledge is hard to acquire, it hinders the flexibility to quickly design and tune digital synthesizers for diverse sounds. In this paper, we propose ``NAS-FM'', which adopts neural architecture search (NAS) to build a differentiable frequency modulation (FM) synthesizer. Tunable synthesizers with interpretable controls can be developed automatically from sounds without any prior expert knowledge and manual operating costs. In detail, we train a supernet with a specifically designed search space, including predicting the envelopes of carriers and modulators with different frequency ratios. An evolutionary search algorithm with adaptive oscillator size is then developed to find the optimal relationship between oscillators and the frequency ratio of FM. Extensive experiments on recordings of different instrument sounds show that our algorithm can build a synthesizer fully automatically, achieving better results than handcrafted synthesizers. Audio samples are available at https://nas-fm.github.io/

This paper targets the perceptual task of separating the different interacting voices, i.e., monophonic melodic streams, in a polyphonic musical piece. We target symbolic music, where notes are explicitly encoded, and model this task as a Multi-Trajectory Tracking (MTT) problem from discrete observations, i.e., notes in a pitch-time space. Our approach builds a graph from a musical piece, by creating one node for every note, and separates the melodic trajectories by predicting a link between two notes if they are consecutive in the same voice/stream. This kind of local, greedy prediction is made possible by node embeddings created by a heterogeneous graph neural network that can capture inter- and intra-trajectory information. Furthermore, we propose a new regularization loss that encourages the output to respect the MTT premise of at most one incoming and one outgoing link for every node, favoring monophonic (voice) trajectories; this loss function might also be useful in other general MTT scenarios. Our approach does not use domain-specific heuristics, is scalable to longer sequences and a higher number of voices, and can handle complex cases such as voice inversions and overlaps. We reach new state-of-the-art results for the voice separation task on classical music of different styles. All code, data, and pretrained models are available on https://github.com/manoskary/vocsep_ijcai2023

As we move closer to real-world social AI systems, AI agents must be able to deal with multiparty (group) conversations. Recognizing and interpreting multiparty behaviors is challenging, as the system must recognize individual behavioral cues, deal with the complexity of multiple streams of data from multiple people, and recognize the subtle contingent social exchanges that take place amongst group members. To tackle this challenge, we propose the Multiparty-Transformer (Multipar- T), a transformer model for multiparty behavior modeling. The core component of our proposed approach is Crossperson Attention, which is specifically designed to detect contingent behavior between pairs of people. We verify the effectiveness of Multipar-T on a publicly available video-based group engagement detection benchmark, where it outperforms state-of-the-art approaches in average F-1 scores by 5.2% and individual class F-1 scores by up to 10.0%. Through qualitative analysis, we show that our Crossperson Attention module is able to discover contingent behaviors.

Identifying the potential associations among drugs, microbes and diseases is of great significance in exploring the pathogenesis and improving precision medicine. There are plenty of computational methods for pair-wise association prediction, such as drug-microbe and microbe-disease associations, but few methods focus on the higher-order triple-wise drug-microbe-disease (DMD) associations. Driven by the advancement of hypergraph neural networks (HGNNs), we expect them to fully capture high-order interaction patterns behind the hypergraph formulated by DMD associations and realize sound prediction performance. However, the confirmed DMD associations are insufficient due to the high cost of in vitro screening, which forms a sparse DMD hypergraph and thus brings in suboptimal generalization ability. To mitigate the limitation, we propose a Multi-view Contrastive Learning Hypergraph Neural Network, named MCHNN, for DMD association prediction. We design a novel multi-view contrastive learning on the DMD hypergraph as an auxiliary task, which guides the HGNN to learn more discriminative representations and enhances the generalization ability. Extensive computational experiments show that MCHNN achieves satisfactory performance in DMD association prediction and, more importantly, demonstrate the effectiveness of our devised multi-view contrastive learning on the sparse DMD hypergraph.

Multi-objective search can be used to model many real-world problems that require finding Pareto optimal paths from a specified start state to a specified goal state, while considering different costmetrics such as distance, time, and fuel. The performance of multi-objective search can be improved by making dominance checking—an operation necessary to determine whether or not a path dominates another—more efficient. This was shown in practice by BOA*, a state-of-the-art bi-objective search algorithm, which outperforms previously existing bi-objective search algorithms in part because it adopts a lazy approach towards dominance checking. EMOA*, a recent multi-objective search algorithm, generalizes BOA* to more-than-two objectives using AVL trees for dominance checking. In this paper, we first propose Linear-Time Multi-Objective A* (LTMOA*), an multi-objective search algorithm that implements a more efficient dominance checking than EMOA* using simple data structures like arrays. We then propose an even lazier approach towards dominance checking, and the resulting algorithm, LazyLTMOA*, distinguishes from EMOA* and LTMOA* by removing the dominance checking during node generation. Our experimental results show that LazyLTMOA* outperforms EMOA* by up to an order of magnitude in terms of runtime.