IJCAI 2023 Papers

The context about trips and users from mobility data is valuable for mobile service providers to understand their customers and improve their services. Existing inference methods require a large number of labels for training, which is hard to meet in practice. In this paper, we study a more practical yet challenging setting—contextual inference using mobility data with minimal supervision (i.e., a few labels per class and massive unlabeled data). A typical solution is to apply semi-supervised methods that follow a self-training framework to bootstrap a model based on all features. However, using a limited labeled set brings high risk of overfitting to self-training, leading to unsatisfactory performance. We propose a novel collaborative distillation framework STCOLAB. It sequentially trains spatial and temporal modules at each iteration following the supervision of ground-truth labels. In addition, it distills knowledge to the module being trained using the logits produced by the latest trained module of the other modality, thereby mutually calibrating the two modules and combining the knowledge from both modalities. Extensive experiments on two real-world datasets show STCOLAB achieves significantly more accurate contextual inference than various baselines.

Emotion recognition in conversation, which aims to predict the emotion for all utterances, has attracted considerable research attention in recent years. It is a challenging task since the recognition of the emotion in one utterance involves many complex factors, such as the conversational context, the speaker's background, and the subtle difference between emotion labels. In this paper, we propose a novel framework which mimics the thinking process when modeling these factors. Specifically, we first comprehend the conversational context with a history-oriented prompt to selectively gather information from predecessors of the target utterance. We then model the speaker's background with an experience-oriented prompt to retrieve the similar utterances from all conversations. We finally differentiate the subtle label semantics with a paraphrasing mechanism to elicit the intrinsic label related knowledge. We conducted extensive experiments on three benchmarks. The empirical results demonstrate the superiority of our proposed framework over the state-of-the-art baselines.

Text classification is a fundamental task for natural language processing, and adapting text classification models across domains has broad applications. Self-training generates pseudo-examples from the model's predictions and iteratively trains on the pseudo-examples, i.e., minimizes the loss on the source domain and the Gibbs entropy on the target domain. However, Gibbs entropy is sensitive to prediction errors, and thus, self-training tends to fail when the domain shift is large. In this paper, we propose Meta-Tsallis Entropy minimization (MTEM). MTEM uses an instance adaptive Tsallis entropy to replace the Gibbs entropy and a meta-learning algorithm to optimize the instance adaptive Tsallis entropy on the target domain. To reduce the computation cost of MTEM, we propose an approximation technique to approximate the second-order derivation involved in the meta-learning. To efficiently generate pseudo labels, we propose an annealing sampling mechanism for exploring the model's prediction probability. Theoretically, we prove the convergence of the meta-learning algorithm in MTEM and analyze the effectiveness of MTEM in achieving domain adaptation. Experimentally, MTEM improves the adaptation performance of BERT with an average of 4 percent on the benchmark dataset.

Model-based diagnosis is a principled and broadly applicable AI-based approach to tackle debugging problems in a wide range of areas including software, knowledge bases, circuits, cars, and robots. Whenever the sound and complete computation of fault explanations in a given preference order (e.g., cardinality or probability) is required, all existing diagnosis algorithms suffer from an exponential space complexity. This can prevent their application on memory-restricted devices and for memory-intensive problem cases. As a remedy, we propose RBF-HS, a diagnostic search based on Korf’s seminal RBFS algorithm which can enumerate an arbitrary fixed number of fault explanations in best-first order within linear space bounds, without sacrificing other desirable properties. Evaluations on real-world diagnosis cases show that RBF-HS, when used to compute minimum-cardinality fault explanations, in most cases saves substantial space while requiring only reasonably more or even less time than Reiter’s HS-Tree, one of the most influential diagnostic algorithms with the same properties.

In rank aggregation, members of a population rank issues to decide which are collectively preferred. We focus instead on identifying divisive issues that express disagreements among the preferences of individuals. We analyse the properties of our divisiveness measures and their relation to existing notions of polarisation. We also study their robustness under incomplete preferences and algorithms for control and manipulation of divisiveness. Our results advance our understanding of how to quantify disagreements in collective decision-making.

We introduce new power indices to measure the a priori voting power of voters in liquid democracy elections where an underlying network restricts delegations. We argue that our power indices are natural extensions of the standard Penrose-Banzhaf index in simple voting games. We show that computing the criticality of a voter is #P-hard even in weighted games with weights polynomially-bounded in the size of the instance. However, for specific settings, such as when the underlying network is a bipartite or complete graph, recursive formulas can compute these indices for weighted voting games in pseudo-polynomial time. We highlight their theoretical properties and provide numerical results to illustrate how restricting the possible delegations can alter voters' voting power.

As humans, we hear sound every second of our life. The sound we hear is often affected by the acoustics of the environment surrounding us. For example, a spacious hall leads to more reverberation. Room Impulse Responses (RIR) are commonly used to characterize environment acoustics as a function of the scene geometry, materials, and source/receiver locations. Traditionally, RIRs are measured by setting up a loudspeaker and microphone in the environment for all source/receiver locations, which is time-consuming and inefficient. We propose to let two robots measure the environment's acoustics by actively moving and emitting/receiving sweep signals. We also devise a collaborative multi-agent policy where these two robots are trained to explore the environment's acoustics while being rewarded for wide exploration and accurate prediction. We show that the robots learn to collaborate and move to explore environment acoustics while minimizing the prediction error. To the best of our knowledge, we present the very first problem formulation and solution to the task of collaborative environment acoustics measurements with multiple agents.

Keeping risk under control is often more crucial than maximizing expected rewards in real-world decision-making situations, such as finance, robotics, autonomous driving, etc. The most natural choice of risk measures is variance, while it penalizes the upside volatility as much as the downside part. Instead, the (downside) semivariance, which captures negative deviation of a random variable under its mean, is more suitable for risk-averse proposes. This paper aims at optimizing the mean-semivariance (MSV) criterion in reinforcement learning w.r.t. steady reward distribution. Since semivariance is time-inconsistent and does not satisfy the standard Bellman equation, the traditional dynamic programming methods are inapplicable to MSV problems directly. To tackle this challenge, we resort to Perturbation Analysis (PA) theory and establish the performance difference formula for MSV. We reveal that the MSV problem can be solved by iteratively solving a sequence of RL problems with a policy-dependent reward function. Further, we propose two on-policy algorithms based on the policy gradient theory and the trust region method. Finally, we conduct diverse experiments from simple bandit problems to continuous control tasks in MuJoCo, which demonstrate the effectiveness of our proposed methods.

Consider oriented graph nodes requiring periodic visits by a service agent. The agent moves among the nodes and receives a payoff for each completed service task, depending on the time elapsed since the previous visit to a node. We consider the problem of finding a suitable schedule for the agent to maximize its long-run average payoff per time unit. We show that the problem of constructing an epsilon-optimal schedule is PSPACE-hard for every fixed non-negative epsilon, and that there exists an optimal periodic schedule of exponential length. We propose randomized finite-memory (RFM) schedules as a compact description of the agent's strategies and design an efficient algorithm for constructing RFM schedules. Furthermore, we construct deterministic periodic schedules by sampling from RFM schedules.

The real-world deployment of fair allocation algorithms usually involves a heterogeneous population of users, which makes it challenging for the users to get complete knowledge of the allocation except for their own bundles. Recently, a new fairness notion, maximin-awareness (MMA) was proposed and it guarantees that every agent is not the worst-off one, no matter how the items that are not allocated to this agent are distributed. We adapt and generalize this notion to the case of indivisible chores and when the agents may have arbitrary weights. Due to the inherent difficulty of MMA, we also consider its up to one and up to any relaxations. A string of results on the existence and computation of MMA related fair allocations, and their connections to existing fairness concepts is given.

In this paper, we introduce Max Markov Chain (MMC), a novel model for sequential data with sparse correlations among the state variables. It may also be viewed as a special class of approximate models for High-order Markov Chains (HMCs). MMC is desirable for domains where the sparse correlations are long-term and vary in their temporal stretches. Although generally intractable, parameter optimization for MMC can be solved analytically. However, based on this result, we derive an approximate solution that is highly efficient empirically. When compared with HMC and approximate HMC models, MMC combines better sample efficiency, model parsimony, and an outstanding computational advantage. Such a quality allows MMC to scale to large domains where the competing models would struggle to perform. We compare MMC with several baselines with synthetic and real-world datasets to demonstrate MMC as a valuable alternative for stochastic modeling.

Image matting is a fundamental technique in visual understanding and has become one of the most significant capabilities in mobile phones. Despite the development of mobile storage and computing power, achieving diverse mobile Artificial Intelligence Generated Content (AIGC) applications remains a great challenge. To address this issue, we present an innovative demonstration of an automatic system called "Matting Moments" that enables automatic image editing based on matting models in different scenarios. Coupled with accurate and refined matting subjects, our system provides visual element editing abilities and backend services for distribution and recommendation that respond to emotional expressions. Our system comprises three components: 1) photo content structuring, 2) data-driven matting engine, and 3) AIGC functions for generation, which automatically achieve diverse photo beautification in the gallery. This system offers a unified framework that guides consumers to obtain intelligent recommendations with beautifully generated contents, helping them enjoy the moments and memories of their present life.

Assigning applicants to posts in the presence of the preferences of applicants and quotas associated with posts is extensively investigated. For a post, lower quota guarantees, and upper quota limits the number of applicants assigned to it. Typically, quotas are assumed to be fixed, which need not be the case in practice. We address this by introducing a soft quota setting, in which every post is associated with two values – lower target and upper target which together denote a range for the intended number of applicants in any assignment. Unlike the fixed quota setting, we allow the number of applicants assigned to a post to fall outside the range. This leads to assignments with deviation. Here, we study the problem of computing an assignment that has two orthogonal optimization objectives – minimizing the deviation (maximum or total) w.r.t. soft quotas and ensuring optimality w.r.t. preferences of applicants (rank-maximality or fairness). The order in which these objectives are considered, the different possibilities to optimize deviation combined with the well-studied notions of optimality w.r.t. preferences open up a range of optimization problems of practical importance. We present efficient algorithms based on flow-networks to solve these optimization problems.

Domain gap between synthetic and real data in visual regression (e.g., 6D pose estimation) is bridged in this paper via global feature alignment and local refinement on the coarse classification of discretized anchor classes in target space, which imposes a piece-wise target manifold regularization into domain-invariant representation learning. Specifically, our method incorporates an explicit self-supervised manifold regularization, revealing consistent cumulative target dependency across domains, to a self-training scheme (e.g., the popular Self-Paced Self-Training) to encourage more discriminative transferable representations of regression tasks. Moreover, learning unified implicit neural functions to estimate relative direction and distance of targets to their nearest class bins aims to refine target classification predictions, which can gain robust performance against inconsistent feature scaling sensitive to UDA regressors. Experiment results on three public benchmarks of the challenging 6D pose estimation task can verify the effectiveness of our method, consistently achieving superior performance to the state-of-the-art for UDA on 6D pose estimation. Codes and pre-trained models are available https://github.com/Gorilla-Lab-SCUT/MAST.

We survey recent work on machine learning (ML) techniques for selecting cutting planes (or cuts) in mixed-integer linear programming (MILP). Despite the availability of various classes of cuts, the task of choosing a set of cuts to add to the linear programming (LP) relaxation at a given node of the branch-and-bound (B&B) tree has defied both formal and heuristic solutions to date. ML offers a promising approach for improving the cut selection process by using data to identify promising cuts that accelerate the solution of MILP instances. This paper presents an overview of the topic, highlighting recent advances in the literature, common approaches to data collection, evaluation, and ML model architectures. We analyze the empirical results in the literature in an attempt to quantify the progress that has been made and conclude by suggesting avenues for future research.

This paper explores how machine learning can help classify aid activities by sector using the OECD Creditor Reporting System (CRS). The CRS is a key source of data for monitoring and evaluating aid flows in line with the United Nations Sustainable Development Goals (SDGs), especially SDG17 which calls for global partnership and data sharing. To address the challenges of current labor-intensive practices of assigning the code and the related human inefficiencies, we propose a machine learning solution that uses ELECTRA to suggest relevant five-digit purpose codes in CRS for aid activities, achieving an accuracy of 0.9575 for the top-3 recommendations. We also conduct qualitative research based on semi-structured interviews and focus group discussions with SDG experts who assess the model results and provide feedback. We discuss the policy, practical, and methodological implications of our work and highlight the potential of AI applications to improve routine tasks in the public sector and foster partnerships for achieving the SDGs.

Modern data processing applications often combine information from a variety of complex sources. Oftentimes, some of these sources, like Machine-Learning systems or crowd-sourced data, are not strictly binary but associated with some degree of confidence in the observation. Ideally, reasoning over such data should take this additional information into account as much as possible. To this end, we propose extensions of Datalog and Datalog+/- to the semantics of Lukasiewicz logic Ł, one of the most common fuzzy logics. We show that such an extension preserves important properties from the classical case and how these properties can lead to efficient reasoning procedures for these new languages.

Pansharpening is to fuse a panchromatic (PAN) image with a multispectral (MS) image to obtain a high-spatial-resolution multispectral (HRMS) image. The deep learning-based pansharpening methods usually apply the convolution operation to extract features and only consider the similarity of gradient information between PAN and HRMS images, resulting in the problems of edge blur and spectral distortion in the fusion results. To solve this problem, a multi-supervised mask protection network (MMPN) is proposed to prevent spatial information from being damaged and overcome spectral distortion in the learning process. Firstly, by analyzing the relationships between high-resolution images and corresponding degraded images, a mask protection strategy (MPS) for edge protection is designed to guide the recovery of fused images. Then, based on the MPS, an MMPN containing four branches is constructed to generate the fusion and mask protection images. In MMPN, each branch employs a dual-stream multi-scale feature fusion module (DMFFM), which is built to extract and fuse the features of two input images. Finally, different loss terms are defined for the four branches, and combined into a joint loss function to realize network training. Experiments on simulated and real satellite datasets show that our method is superior to state-of-the-art methods both subjectively and objectively.

The visual quality of point clouds has been greatly emphasized since the ever-increasing 3D vision applications are expected to provide cost-effective and high-quality experiences for users. Looking back on the development of point cloud quality assessment (PCQA), the visual quality is usually evaluated by utilizing single-modal information, i.e., either extracted from the 2D projections or 3D point cloud. The 2D projections contain rich texture and semantic information but are highly dependent on viewpoints, while the 3D point clouds are more sensitive to geometry distortions and invariant to viewpoints. Therefore, to leverage the advantages of both point cloud and projected image modalities, we propose a novel no-reference Multi-Modal Point Cloud Quality Assessment (MM-PCQA) metric. In specific, we split the point clouds into sub-models to represent local geometry distortions such as point shift and down-sampling. Then we render the point clouds into 2D image projections for texture feature extraction. To achieve the goals, the sub-models and projected images are encoded with point-based and image-based neural networks. Finally, symmetric cross-modal attention is employed to fuse multi-modal quality-aware information. Experimental results show that our approach outperforms all compared state-of-the-art methods and is far ahead of previous no-reference PCQA methods, which highlights the effectiveness of the proposed method. The code is available at https://github.com/zzc-1998/MM-PCQA.