IJCAI 2024 论文列表

Search in test time is often used to improve the performance of reinforcement learning algorithms. Performing theoretically sound search in fully adversarial two-player games with imperfect information is notoriously difficult and requires a complicated training process. We present a method for adding test-time search to an arbitrary policy-gradient algorithm that learns from sampled trajectories. Besides the policy network, the algorithm trains an additional critic network, which estimates the expected values of players following various transformations of the policies given by the policy network. These values are then used for depth-limited search. We show how the values from this critic can create a value function for imperfect information games. Moreover, they can be used to compute the summary statistics necessary to start the search from an arbitrary decision point in the game. The presented algorithm is scalable to very large games since it does not require any search during train time. We evaluate the algorithm's performance when trained along Regularized Nash Dynamics, and we evaluate the benefit of using the search in the standard benchmark game of Leduc hold'em, multiple variants of imperfect information Goofspiel, and Battleships.

Transformer models have been successful in various sequence processing tasks, but the self-attention mechanism's computational cost limits its practicality for long sequences. Although there are existing attention variants that improve computational efficiency, they have a limited ability to abstract global information effectively based on their hand-crafted mixing strategies. On the other hand, state-space models (SSMs) are tailored for long sequences but cannot capture complicated local information. Therefore, the combination of them as a unified token mixer is a trend in recent long-sequence models. However, the linearized attention degrades performance significantly even when equipped with SSMs. To address the issue, we propose a new method called LongVQ. LongVQ uses the vector quantization (VQ) technique to compress the global abstraction as a length-fixed codebook, enabling the linear-time computation of the attention matrix. This technique effectively maintains dynamic global and local patterns, which helps to complement the lack of long-range dependency issues. Our experiments on the Long Range Arena benchmark, autoregressive language modeling, and image and speech classification demonstrate the effectiveness of LongVQ. Our model achieves significant improvements over other sequence models, including variants of Transformers, Convolutions, and recent State Space Models.

Urban villages are areas filled with rural-like improvised structures in Chinese cities, usually housing the most vulnerable groups. Under the guidance of the Sustainable Development Goals (SDGs), the Chinese government initiated renewal and redevelopment projects, underscoring the meticulous mapping and segmentation of urban villages. Satellite imagery is advanced and efficient in identifying urban villages and monitoring changes, but traditional methods neglect the morphological diversity in season, shape, size, spacing, and layout of urban villages, which is not satisfying for long-term wide-range data. Here, we design a targeted approach based on Tobler’s First Law of Geography, using curriculum labeling to solve morphological diversity and semi-automatically generate segmentation for urban village boundaries. Specifically, we use manually labeled data as seeds for pre-trained SegFormer models and incrementally fine-tune the model based on geographical proximity. The rigorous experimentation across five diverse cities substantiates the commendable efficacy of our methodology. IoU metric demonstrates a noteworthy improvement of over 119% to baseline. Our final results cover 265,050 urban villages across 433 cities in China over the past 10 years, and the analysis reveals the uneven redevelopment by geography and city scale. We further examine the within-city distribution and verify the urban scaling law associated with several socio-economic factors. Our method can be used nationwide to decide redevelopment priority and resource tilt, contributing to SDG 11.1 on affordable housing and upgrading slums. The code and dataset are available at https://github.com/tsinghua-fib-lab/LtCUV.

Video anomaly detection (VAD) is the core problem of intelligent video surveillance. Previous methods commonly adopt the unsupervised paradigm of frame reconstruction or prediction. However, the lack of mining of temporal dependent relationships and diversified event patterns within videos limit the performance of existing methods. To tackle these problems, we propose a novel prototype-guided and dynamic-aware long-distance frame prediction paradigm for VAD. Specifically, we develop a prototype-guided dynamics matching network (PDM-Net) to enhance the discriminant and robustness of anomaly detector. To explore the temporal contexts, we equip PDM-Net with a long short-term dynamic prototype alignment learning mechanism, which stores long-term dynamic prototypes into memory bank and learns how to recall long-term dynamic prototypes with short-term dynamics. As a result, the short input sequences can recall long-term dynamic prototypes stored in the memory bank to achieve the task of long-distance frame prediction. Besides, a feature discrimination module is adopted to extract the representative dynamic features of various normal events meanwhile preserving the diversity of normal patterns. Experimental results on four public datasets demonstrate the superiority of our method.

The Mixtures-of-Experts (MoE) model is a widespread distributed and integrated learning method for large language models (LLM), which is favored due to its ability to sparsify and expand models efficiently. However, the performance of MoE is limited by load imbalance and high latency of All-to-All communication, along with relatively redundant computation owing to large expert capacity. Load imbalance may result from existing routing policies that consistently tend to select certain experts. The frequent inter-node communication in the All-to-All procedure also significantly prolongs the training time. To alleviate the above performance problems, we propose a novel routing strategy that combines load balance and locality by converting partial inter-node communication to that of intra-node. Notably, we elucidate that there is a minimum threshold for expert capacity, calculated through the maximal angular deviation between the gating weights of the experts and the assigned tokens. We port these modifications on the PanGu-Σ model based on the MindSpore framework with multi-level routing and conduct experiments on Ascend clusters. The experiment results demonstrate that the proposed LocMoE reduces training time per epoch by 12.68% to 22.24% compared to classical routers, such as hash router and switch router, without impacting the model accuracy.

Spiking Neural Networks (SNNs) mimic the information-processing mechanisms of the human brain and are highly energy-efficient, making them well-suited for low-power edge devices. However, the pursuit of accuracy in current studies leads to large, long-timestep SNNs, conflicting with the resource constraints of these devices. In order to design lightweight and efficient SNNs, we propose a new approach named LitE-SNN that incorporates both spatial and temporal compression into the automated network design process. Spatially, we present a novel Compressive Convolution block (CompConv) to expand the search space to support pruning and mixed-precision quantization. Temporally, we are the first to propose a compressive timestep search to identify the optimal number of timesteps under specific computation cost constraints. Finally, we formulate a joint optimization to simultaneously learn the architecture parameters and spatial-temporal compression strategies to achieve high performance while minimizing memory and computation costs. Experimental results on CIFAR-10, CIFAR-100, and Google Speech Command datasets demonstrate our proposed LitE-SNNs can achieve competitive or even higher accuracy with remarkably smaller model sizes and fewer computation costs.

The railway scheduling problem requires the computation of an operable timetable that satisfies constraints involving railway infrastructure and resource occupancy times, while minimising average delay over a set of events. Since this problem is computationally hard, practical solutions typically roll out feasible (but suboptimal) schedules one step at a time, by choosing which train to move next in every step. The choices made by such algorithms are necessarily myopic, and incur the risk of driving the system to a deadlock. To escape deadlocks, the predominant approach is to stay away from states flagged as potentially unsafe by some fast-to-compute rule R. While many choices of R guarantee deadlock avoidance, they are suboptimal in the sense of also flagging some safe states as unsafe. In this paper, we revisit the literature on process scheduling and describe a rule R0 that is (i) necessary and sufficient for deadlock detection when the network has at least two tracks in each resource (station / track section), (ii) computable in linear time, and (iii) yields lower delays when combined with existing scheduling algorithms on both synthetic and real data sets from Indian Railways.

Lifted planning is usually defined as planning directly over a first-order representation. From the mid-1990s until the late 2010s, lifted planning was sidelined, as most of the state-of-the-art planners first ground the task and then solve it using a propositional representation. Moreover, it was unclear whether lifted planners could scale. But as planning problems become harder, they also become infeasible to ground. Recently, lifted planners came back into play, aiming at problems where grounding is a bottleneck. In this work, we survey recent advances in lifted planning. The main techniques rely either on state-space search or logic satisfiability. For lifted search-based planners, we show the direct connections to other areas of computer science, such as constraint satisfaction problems and databases. For lifted planners based on satisfiability, the advances in modeling are crucial to their scalability. We briefly describe the main planners available in the literature and their techniques.

Generalized Category Discovery (GCD) presents a realistic and challenging problem in open-world learning. Given a par- tially labeled dataset, GCD aims to categorize unlabeled data by leveraging visual knowledge from the labeled data, where the unlabeled data includes both known and unknown classes. Existing methods based on parametric/non-parametric classi- fiers attempt to generate pseudo-labels/relationships for the unlabeled data to enhance representation learning. However, the lack of ground-truth labels for novel classes often leads to noisy pseudo-labels/relationships, resulting in suboptimal representation learning. This paper introduces a novel method using Nearest Neighbor Distance-aware Label Consistency sample selection. It creates class-consistent subsets for novel class sample clusters from the current GCD method, acting as “pseudo-labeled sets” to mitigate representation bias. We propose progressive supervised representation learning with selected samples to optimize the trade-off between quantity and purity in each subset. Our method is versatile and appli- cable to various GCD methods, whether parametric or non- parametric. We conducted extensive experiments on multiple generic and fine-grained image classification datasets to eval- uate the effectiveness of our approach. The results demon- strate the superiority of our method in achieving improved performance in generalized category discovery tasks.

Existing efforts on human tracking using wireless signal are primarily focused on constrained scenarios with only a few individuals in empty spaces. However, in practical unconstrained scenarios with severe interference and attenuation, accurate multi-person tracking has been intractable. In this paper, we propose NeuralTBD, utilizing the capability of deep models and advancement of Tracking-Before-Detect (TBD) methodology to achieve accurate human tracking. TBD is a classical tracking methodology from signal processing accumulating measurement in time domain to distinguish target traces from interference, which however relies on handcrafted shape/motion models, impeding efficacy in complex indoor scenarios. To tackle this challenge, we build an end-to-end learning-based TBD framework leverages the advanced modeling capabilities of deep models to significantly enhance the performance of TBD. To evaluate NeuralTBD, we collect an RF-based tracking dataset in unconstrained scenarios, which encompasses 4 million annotated radar frames with up to 19 individuals acting in 6 different scenarios. NeuralTBD realizes a 70% improvement in performance compared to conventional TBD methods. To our knowledge, this is the first attempt dealing with RF-based unconstrained human tracking. The code and dataset will be released.

This paper discusses the revenue management (RM) problem to maximize revenue by pricing items or services. One challenge in this problem is that the demand distribution is unknown and varies over time in real applications such as airline and retail industries. In particular, the time-varying demand has not been well studied under scenarios of unknown demand due to the difficulty of jointly managing the remaining inventory and estimating the demand. To tackle this challenge, we first introduce an episodic generalization of the RM problem motivated by typical application scenarios. We then propose a computationally efficient algorithm based on posterior sampling, which effectively optimizes prices by solving linear programming. We derive a Bayesian regret upper bound of this algorithm for general models where demand parameters can be correlated between time periods, while also deriving a regret lower bound for generic algorithms. Our empirical study shows that the proposed algorithm performs better than other benchmark algorithms and comparably to the optimal policy in hindsight. We also propose a heuristic modification of the proposed algorithm, which further efficiently learns the pricing policy in the experiments. An extended version of this paper with appendixes is available at: http://arxiv.org/abs/2405.04910.

Geometry Problem Solving (GPS), which is a classic and challenging math problem, has attracted much attention in recent years. It requires a solver to comprehensively understand both text and diagram, master essential geometry knowledge, and appropriately apply it in reasoning. However, existing works follow a paradigm of neural machine translation and only focus on enhancing the capability of encoders, which neglects the essential characteristics of human geometry reasoning. In this paper, inspired by dual-process theory, we propose a Dual-Reasoning Geometry Solver (DualGeoSolver) to simulate the dual-reasoning process of humans for GPS. Specifically, we construct two systems in DualGeoSolver, namely Knowledge System and Inference System. Knowledge System controls an implicit reasoning process, which is responsible for providing diagram information and geometry knowledge according to a step-wise reasoning goal generated by Inference System. Inference System conducts an explicit reasoning process, which specifies the goal in each reasoning step and applies the knowledge to generate program tokens for resolving it. The two systems carry out the above process iteratively, which behaves more in line with human cognition. We conduct extensive experiments on two benchmark datasets, GeoQA and GeoQA+. The results demonstrate the superiority of DualGeoSolver in both solving accuracy and robustness from explicitly modeling human reasoning process and knowledge application.

Social dilemmasare situations of inter-dependent decision making in which individualrationality can lead to outcomes with poor social qualities. The ubiquity of social dilem-mas in social, biological, and computational systems has generated substantial researchacross these diverse disciplines into the study of mechanisms for avoiding deficient outcomes by promoting and maintaining mutual cooperation. Much of this research is focused on studying how individuals faced with a dilemma can learn to cooperate by adapting their behaviours according to their past experience. In particular, three types of learning approaches have been studied: evolutionary game-theoretic learning, reinforcement learning, and best-response learning. This article is a comprehensive integrated survey of these learning approaches in the context of dilemma games. We formally introduce dilemma games and their inherent challenges. We then outline the three learning approaches and, for eachapproach, provide a survey of the solutions proposed for dilemma resolution. Finally, we provide a comparative summary and discuss directions in which further research is needed.

Three-dimensional Ground Penetrating Radar (3D GPR) data offer comprehensive views of the subsurface, yet identifying and classifying underground anomalies from this data is challenging due to limitations like scarce training data and variable underground environments. In response, we introduce learning in the Cube Reservoir Network (CubeRes) model space for efficient and accurate subsurface anomaly detection. CubeRes, incorporating three reservoirs, captures the dynamics in both horizontal and vertical directions inherent in the 3D GPR data. Fitting the data with CubeRes, representing the data with the compact fitted model, and measuring the difference between models by a proposed parameterized model metric, the original data is transformed from the data space to the CubeRes model space. Subsequently, we introduce an optimization strategy in this model space, aimed at bolstering fitting accuracy and improving category discrimination. This enhancement facilitates a more nuanced differentiation of dynamics across various GPR data categories, thereby enabling effective classification on the models rather than the original data. Experiments on real-world data validate our method's effectiveness and superiority, particularly in data-limited scenarios.

The success of deep learning depends on large-scale and well-curated training data, while data in real-world applications are commonly long-tailed and noisy. Existing methods are usually dependent on label frequency to tackle class imbalance, while the model bias on different classes is not directly related to label frequency and the true label frequency is inaccessible under label noise. To solve this, we propose a robust method for learning from long-tailed noisy data with sample selection and balanced loss. Specifically, we separate the noisy training data into clean labeled set and unlabeled set with sample selection, and train the deep neural network in a semi-supervised manner with a balanced loss based on model bias. Extensive experiments on benchmarks demonstrate that our method outperforms existing state-of-the-art methods.

Recently, spiking neural networks (SNNs) have demonstrated substantial potential in computer vision tasks. In this paper, we present an Efficient Spiking Deraining Network, called ESDNet. Our work is motivated by the observation that rain pixel values will lead to a more pronounced intensity of spike signals in SNNs. However, directly applying deep SNNs to image deraining task still remains a significant challenge. This is attributed to the information loss and training difficulties that arise from discrete binary activation and complex spatiotemporal dynamics. To this end, we develop a spiking residual block to convert the input into spike signals, then adaptively optimize the membrane potential by introducing attention weights to adjust spike responses in a data-driven manner, alleviating information loss caused by discrete binary activation. By this way, our ESDNet can effectively detect and analyze the characteristics of rain streaks by learning their fluctuations. This also enables better guidance for the deraining process and facilitates high-quality image reconstruction. Instead of relying on the ANN-SNN conversion strategy, we introduce a gradient proxy strategy to directly train the model for overcoming the challenge of training. Experimental results show that our approach gains comparable performance against ANN-based methods while reducing energy consumption by 54%. The code source is available at https://github.com/MingTian99/ESDNet.

Monitoring is a runtime verification technique that can be used to check whether an execution of a system (trace) satisfies or not a given set of properties. Compared to other formal verification techniques, e.g., model checking, one needs to specify the properties to be monitored, but a complete model of the system is no longer necessary. First, we introduce the pure past fragment of Signal Temporal Logic (ppSTL), and we use it to define the monitorable safety (G(ppSTL)) and cosafety (F(ppSTL)) fragments of STL, which properly extend the commonly-used bounded-future fragment. Then, we devise a multi-objective genetic programming algorithm to automatically extend the set of properties to monitor on the basis of the history of failure traces collected over time. The framework resulting from the integration of the monitor and the learning algorithm is then experimentally validated on various public datasets. The outcomes of the experimentation confirm the effectiveness of the proposed solution.

In Emergent Communication (EC) agents learn to communicate with one another, but the protocols that they develop are specialised to their training community. This observation led to research into Zero-Shot Coordination (ZSC) for learning communication strategies that are robust to agents not encountered during training. However, ZSC typically assumes that no prior data is available about the agents that will be encountered in the zero-shot setting. In many cases, this presents an unnecessarily hard problem and rules out communication via preestablished conventions. We propose a novel AI challenge called a Cooperative Language Acquisition Problem (CLAP) in which the ZSC assumptions are relaxed by allowing a 'joiner' agent to learn from a dataset of interactions between agents in a target community. We propose and compare two methods for solving CLAPs: Behaviour Cloning (BC), and Emergent Communication pretraining and Translation Learning (ECTL), in which an agent is trained in self-play with EC and then learns to translate between an emergent protocol and the target community's protocol.

This paper provides a comprehensive review of deep structural causal models (DSCMs), particularly focusing on their ability to answer counterfactual queries using observational data within known causal structures. It delves into the characteristics of DSCMs by analyzing the hypotheses, guarantees, and applications inherent to the underlying deep learning components and structural causal models, fostering a finer understanding of their capabilities and limitations in addressing different counterfactual queries. Furthermore, it highlights the challenges and open questions in the field of deep structural causal modeling. It sets the stages for researchers to identify future work directions and for practitioners to get an overview in order to find out the most appropriate methods for their needs.