Shuyue Hu

Ziyu Wan, Yunxiang Li, Xiaoyu Wen et al.

Recent research on Reasoning of Large Language Models (LLMs) has sought to further enhance their performance by integrating meta-thinking -- enabling models to monitor, evaluate, and control their reasoning processes for more adaptive and effective problem-solving. However, current single-agent work lacks a specialized design for acquiring meta-thinking, resulting in low efficacy. To address this challenge, we introduce Reinforced Meta-thinking Agents (ReMA), a novel framework that leverages Multi-Agent Reinforcement Learning (MARL) to elicit meta-thinking behaviors, encouraging LLMs to think about thinking. ReMA decouples the reasoning process into two hierarchical agents: a high-level meta-thinking agent responsible for generating strategic oversight and plans, and a low-level reasoning agent for detailed executions. Through iterative reinforcement learning with aligned objectives, these agents explore and learn collaboration, leading to improved generalization and robustness. Empirical results from single-turn experiments demonstrate that ReMA outperforms single-agent RL baselines on complex reasoning tasks, including competitive-level mathematical benchmarks and LLM-as-a-Judge benchmarks. Additionally, we further extend ReMA to multi-turn interaction settings, leveraging turn-level ratio and parameter sharing to improve efficiency. Comprehensive ablation studies further illustrate the evolving dynamics of each distinct agent, providing valuable insights into how the meta-thinking reasoning process enhances the reasoning capabilities of LLMs. Our code can be found in https://github.com/ziyuwan/ReMA-public

Guibin Zhang, Kaijie Chen, Guancheng Wan et al.

The past two years have witnessed the evolution of large language model (LLM)-based multi-agent systems from labor-intensive manual design to partial automation (\textit{e.g.}, prompt engineering, communication topology) and eventually to fully automated design. However, existing agentic automation pipelines often lack LLM heterogeneity and focus on single-objective performance optimization, limiting their potential to combine weaker models for more customized and cost-effective solutions. To address this challenge, we propose EvoFlow, a niching evolutionary algorithm-based framework to automatically search a population of heterogeneous and complexity-adaptive agentic workflows, rather than a single homogeneous, complex workflow. Technically, EvoFlow performs \textit{(1) tag-based retrieval} to extract parent workflows from an agentic population, evolves new workflows through \textit{(2) crossover} and \textit{(3) mutation}, and employs \textit{(4) niching-based selection} to maintain population diversity and quality. Extensive evaluations across seven benchmarks demonstrate that EvoFlow is: \textbf{(I) diverse}, evolving a population of workflows ranging from simple I/O tasks to complex multi-turn interactions; \textbf{(II) high-performing}, outperforming previous handcrafted and automated workflows by $1.23\%\sim29.86\%$; \textbf{(III) economical}, surpassing powerful \llmname{o1-preview} at $12.4\%$ of its inference cost using weaker open-source models.

Chenxu Wang, Hao Li, Yiqun Zhang et al.

Large language models (LLMs) often exhibit complementary strengths. Model routing harnesses these strengths by dynamically directing each query to the most suitable model, given a candidate model pool. However, routing performance relies on accurate model representations, and adding new models typically requires retraining, limiting scalability. To address these challenges, we propose a novel routing method using in-context vectors to represent model capabilities. The method proceeds in two stages. First, queries are embedded and projected into vectors, with a projector and LLM-based router trained to reconstruct the original queries, aligning vector representations with the router's semantic space. Second, each candidate model is profiled on a query set, and the router learns -- based on in-context vectors of query and model performance -- to predict whether each model can correctly answer new queries. Extensive experiments demonstrate that our method achieves state-of-the-art routing performance in both in-distribution and out-of-distribution tasks. Moreover, our method allows for seamless integration of new models without retraining the router. The code is available at https://github.com/lalalamdbf/ICL-Router.

Chunjiang Mu, Hao Guo, Yang Chen et al.

The study of cooperation within social dilemmas has long been a fundamental topic across various disciplines, including computer science and social science. Recent advancements in Artificial Intelligence (AI) have significantly reshaped this field, offering fresh insights into understanding and enhancing cooperation. This survey examines three key areas at the intersection of AI and cooperation in social dilemmas. First, focusing on multi-agent cooperation, we review the intrinsic and external motivations that support cooperation among rational agents, and the methods employed to develop effective strategies against diverse opponents. Second, looking into human-agent cooperation, we discuss the current AI algorithms for cooperating with humans and the human biases towards AI agents. Third, we review the emergent field of leveraging AI agents to enhance cooperation among humans. We conclude by discussing future research avenues, such as using large language models, establishing unified theoretical frameworks, revisiting existing theories of human cooperation, and exploring multiple real-world applications.

Qiaosheng Zhang, Chenjia Bai, Shuyue Hu et al.

This work designs and analyzes a novel set of algorithms for multi-agent reinforcement learning (MARL) based on the principle of information-directed sampling (IDS). These algorithms draw inspiration from foundational concepts in information theory, and are proven to be sample efficient in MARL settings such as two-player zero-sum Markov games (MGs) and multi-player general-sum MGs. For episodic two-player zero-sum MGs, we present three sample-efficient algorithms for learning Nash equilibrium. The basic algorithm, referred to as MAIDS, employs an asymmetric learning structure where the max-player first solves a minimax optimization problem based on the joint information ratio of the joint policy, and the min-player then minimizes the marginal information ratio with the max-player's policy fixed. Theoretical analyses show that it achieves a Bayesian regret of tilde{O}(sqrt{K}) for K episodes. To reduce the computational load of MAIDS, we develop an improved algorithm called Reg-MAIDS, which has the same Bayesian regret bound while enjoying less computational complexity. Moreover, by leveraging the flexibility of IDS principle in choosing the learning target, we propose two methods for constructing compressed environments based on rate-distortion theory, upon which we develop an algorithm Compressed-MAIDS wherein the learning target is a compressed environment. Finally, we extend Reg-MAIDS to multi-player general-sum MGs and prove that it can learn either the Nash equilibrium or coarse correlated equilibrium in a sample efficient manner.

Die Hu, Shuyue Hu, Chunjiang Mu et al.

Understanding and predicting the behavior of large-scale multi-agents in games remains a fundamental challenge in multi-agent systems. This paper examines the role of heterogeneity in equilibrium formation by analyzing how smooth regret-matching drives a large number of heterogeneous agents with diverse initial policies toward unified behavior. By modeling the system state as a probability distribution of regrets and analyzing its evolution through the continuity equation, we uncover a key phenomenon in diverse multi-agent settings: the variance of the regret distribution diminishes over time, leading to the disappearance of heterogeneity and the emergence of consensus among agents. This universal result enables us to prove convergence to quantal response equilibria in both competitive and cooperative multi-agent settings. Our work advances the theoretical understanding of multi-agent learning and offers a novel perspective on equilibrium selection in diverse game-theoretic scenarios.

Zhongtian Ma, Qiaosheng Zhang, Bocheng Zhou et al.

Despite the growing popularity of graph attention mechanisms, their theoretical understanding remains limited. This paper aims to explore the conditions under which these mechanisms are effective in node classification tasks through the lens of Contextual Stochastic Block Models (CSBMs). Our theoretical analysis reveals that incorporating graph attention mechanisms is \emph{not universally beneficial}. Specifically, by appropriately defining \emph{structure noise} and \emph{feature noise} in graphs, we show that graph attention mechanisms can enhance classification performance when structure noise exceeds feature noise. Conversely, when feature noise predominates, simpler graph convolution operations are more effective. Furthermore, we examine the over-smoothing phenomenon and show that, in the high signal-to-noise ratio (SNR) regime, graph convolutional networks suffer from over-smoothing, whereas graph attention mechanisms can effectively resolve this issue. Building on these insights, we propose a novel multi-layer Graph Attention Network (GAT) architecture that significantly outperforms single-layer GATs in achieving \emph{perfect node classification} in CSBMs, relaxing the SNR requirement from $ ω(\sqrt{\log n}) $ to $ ω(\sqrt{\log n} / \sqrt[3]{n}) $. To our knowledge, this is the first study to delineate the conditions for perfect node classification using multi-layer GATs. Our theoretical contributions are corroborated by extensive experiments on both synthetic and real-world datasets, highlighting the practical implications of our findings.

Yang Chen, Libo Zhang, Jiamou Liu et al.

The recent mean field game (MFG) formalism has enabled the application of inverse reinforcement learning (IRL) methods in large-scale multi-agent systems, with the goal of inferring reward signals that can explain demonstrated behaviours of large populations. The existing IRL methods for MFGs are built upon reducing an MFG to a Markov decision process (MDP) defined on the collective behaviours and average rewards of the population. However, this paper reveals that the reduction from MFG to MDP holds only for the fully cooperative setting. This limitation invalidates existing IRL methods on MFGs with non-cooperative environments. To measure more general behaviours in large populations, we study the use of individual behaviours to infer ground-truth reward functions for MFGs. We propose Mean Field IRL (MFIRL), the first dedicated IRL framework for MFGs that can handle both cooperative and non-cooperative environments. Based on this theoretically justified framework, we develop a practical algorithm effective for MFGs with unknown dynamics. We evaluate MFIRL on both cooperative and mixed cooperative-competitive scenarios with many agents. Results demonstrate that MFIRL excels in reward recovery, sample efficiency and robustness in the face of changing dynamics.

Shuyue Hu, Chin-Wing Leung, Ho-fung Leung et al.

Understanding the evolutionary dynamics of reinforcement learning under multi-agent settings has long remained an open problem. While previous works primarily focus on 2-player games, we consider population games, which model the strategic interactions of a large population comprising small and anonymous agents. This paper presents a formal relation between stochastic processes and the dynamics of independent learning agents who reason based on the reward signals. Using a master equation approach, we provide a novel unified framework for characterising population dynamics via a single partial differential equation (Theorem 1). Through a case study involving Cross learning agents, we illustrate that Theorem 1 allows us to identify qualitatively different evolutionary dynamics, to analyse steady states, and to gain insights into the expected behaviour of a population. In addition, we present extensive experimental results validating that Theorem 1 holds for a variety of learning methods and population games.