The performance of multi-agent reinforcement learning (MARL) in partially observable environments depends on effectively aggregating information from observations, communications, and reward signals. While most existing multi-agent systems primarily rely on rewards as the only feedback for policy training, our research shows that introducing auxiliary predictive tasks can significantly enhance learning efficiency and stability. We propose Belief-based Predictive Auxiliary Learning (BEPAL), a framework that incorporates auxiliary training objectives to support policy optimization. BEPAL follows the centralized training with decentralized execution paradigm. Each agent learns a belief model that predicts unobservable state information, such as other agents' rewards or motion directions, alongside its policy model. By enriching hidden state representations with information that does not directly contribute to immediate reward maximization, this auxiliary learning process stabilizes MARL training and improves overall performance. We evaluate BEPAL in the predator-prey environment and Google Research Football, where it achieves an average improvement of about 16 percent in performance metrics and demonstrates more stable convergence compared to baseline methods.

The recent surge in AI agents that autonomously communicate, collaborate with humans and use diverse tools has unlocked promising opportunities in various real-world settings. However, a vital aspect remains underexplored: how agents handle data. Scalable autonomy demands agents that continuously acquire, process, and evolve their data. In this paper, we argue that data-savvy capabilities should be a top priority in the design of agentic systems to ensure reliable real-world deployment. Specifically, we propose four key capabilities to realize this vision: (1) Proactive data acquisition: enabling agents to autonomously gather task-critical knowledge or solicit human input to address data gaps; (2) Sophisticated data processing: requiring context-aware and flexible handling of diverse data challenges and inputs; (3) Interactive test data synthesis: shifting from static benchmarks to dynamically generated interactive test data for agent evaluation; and (4) Continual adaptation: empowering agents to iteratively refine their data and background knowledge to adapt to shifting environments. While current agent research predominantly emphasizes reasoning, we hope to inspire a reflection on the role of data-savvy agents as the next frontier in data-centric AI.

Translating natural language to SQL remains difficult for complex queries. Such queries often need environmental interaction and self-correction. To address this, we introduce MARS-SQL, a novel multi-agent framework that combines principled task decomposition and interactive reinforcement learning (RL). Our system comprises three specialized agents: a Grounding Agent for schema linking, a Generation Agent for query generation, and a Validation Agent for final selection. The core of our framework is the Generation agent, which is trained via a multi-turn RL policy. Adopting a ReAct-style Think-Act-Observe loop, the agent iteratively generates thoughts, executes SQL actions against a live database, and revises its strategy based on execution feedback, enabling dynamic, stateful reasoning and self-correction. At inference time, we generate multiple interaction trajectories to explore diverse reasoning paths. The Validation agent, then selects the optimal trajectory by modeling verification as a next-token prediction task and choosing the solution with the highest generation probability. This structured workflow pipelines specialized agents. It combines interactive RL for generation with generative modeling for verification. The approach proves highly effective for robust and accurate SQL generation. Experiments show that MARS-SQL achieves state-of-the-art Execution Accuracy of 77.84% on the BIRD dev set and 89.75% on the Spider test set. Our code is available at https://github.com/YangHaolin0526/MARS-SQL.

Recently, effective coordination in embodied multi-agent systems has remained a fundamental challenge, particularly in scenarios where agents must balance individual perspectives with global environmental awareness. Existing approaches often struggle to balance fine-grained local control with comprehensive scene understanding, resulting in limited scalability and compromised collaboration quality. In this paper, we present GauDP, a novel Gaussian-image synergistic representation that facilitates scalable, perception-aware imitation learning in multi-agent collaborative systems. Specifically, GauDP constructs a globally consistent 3D Gaussian field from decentralized RGB observations, then dynamically redistributes 3D Gaussian attributes to each agent's local perspective. This enables all agents to adaptively query task-critical features from the shared scene representation while maintaining their individual viewpoints. This design facilitates both fine-grained control and globally coherent behavior without requiring additional sensing modalities (e.g., 3D point cloud). We evaluate GauDP on the RoboFactory benchmark, which includes diverse multi-arm manipulation tasks. Our method achieves superior performance over existing image-based methods and approaches the effectiveness of point-cloud-driven methods, while maintaining strong scalability as the number of agents increases.

We propose a structure to represent the social fabric of a group. We call it the `cohesion network' of the group. It can be seen as a graph whose vertices are strict subgroups and whose edges indicate a prescribed `pro-social behaviour' from one subgroup towards another. In social psychology, pro-social behaviours are building blocks of full-blown cooperation, which we assimilate here with `group cohesiveness'. We then define a formal framework to study cohesive group agency. To do so, we simply instantiate pro-social behaviour with the more specific relation of `successful assistance' between acting entities in a group. The relations of assistance within a group at the moment of agency constitute the social fabric of the cohesive group agency. We build our logical theory upon the logic of agency "bringing-it-about". We obtain a family of logics of cohesive group agency, one for every class of cohesion networks.

Current benchmarks for coding evaluate language models (LMs) on concrete, well-specified tasks such as fixing specific bugs or writing targeted tests. However, human programmers do not spend all day incessantly addressing isolated tasks. Instead, real-world software development is grounded in the pursuit of high-level goals, like improving user retention or reducing costs. Evaluating whether LMs can also iteratively develop code to better accomplish open-ended objectives without any explicit guidance remains an open challenge. To address this, we introduce CodeClash, a benchmark where LMs compete in multi-round tournaments to build the best codebase for achieving a competitive objective. Each round proceeds in two phases: agents edit their code, then their codebases compete head-to-head in a code arena that determines winners based on objectives like score maximization, resource acquisition, or survival. Whether it's writing notes, scrutinizing documentation, analyzing competition logs, or creating test suites, models must decide for themselves how to improve their codebases both absolutely and against their opponents. We run 1680 tournaments (25,200 rounds total) to evaluate 8 LMs across 6 arenas. Our results reveal that while models exhibit diverse development styles, they share fundamental limitations in strategic reasoning. Models also struggle with long-term codebase maintenance, as repositories become progressively messy and redundant. These limitations are stark: top models lose every round against expert human programmers. We open-source CodeClash to advance the study of autonomous, goal-oriented code development.

This study examines the trade-offs of increasing sampled reasoning paths in self-consistency for modern large language models (LLMs). Earlier research with older models showed that combining multiple reasoning chains improves results before reaching a plateau. Using Gemini 2.5 models on HotpotQA and Math-500, we revisit those claims under current model conditions. Each configuration pooled outputs from varying sampled reasoning paths and compared them to a single chain-of-thought (CoT) baseline. Larger models exhibited a more stable and consistent improvement curve. The results confirm that performance gains taper off after moderate sampling, aligning with past findings. This plateau suggests diminishing returns driven by overlap among reasoning paths. Self-consistency remains useful, but high-sample configurations offer little benefit relative to their computational cost.

Telecom networks are rapidly growing in scale and complexity, making effective management, operation, and optimization increasingly challenging. Although Artificial Intelligence (AI) has been applied to many telecom tasks, existing models are often narrow in scope, require large amounts of labeled data, and struggle to generalize across heterogeneous deployments. Consequently, network troubleshooting continues to rely heavily on Subject Matter Experts (SMEs) to manually correlate various data sources to identify root causes and corrective actions. To address these limitations, we propose a Multi-Agent System (MAS) that employs an agentic workflow, with Large Language Models (LLMs) coordinating multiple specialized tools for fully automated network troubleshooting. Once faults are detected by AI/ML-based monitors, the framework dynamically activates agents such as an orchestrator, solution planner, executor, data retriever, and root-cause analyzer to diagnose issues and recommend remediation strategies within a short time frame. A key component of this system is the solution planner, which generates appropriate remediation plans based on internal documentation. To enable this, we fine-tuned a Small Language Model (SLM) on proprietary troubleshooting documents to produce domain-grounded solution plans. Experimental results demonstrate that the proposed framework significantly accelerates troubleshooting automation across both Radio Access Network (RAN) and Core network domains.

With the rapid progress of large language models (LLMs), LLM-powered multi-agent systems (MAS) are drawing increasing interest across academia and industry. However, many current MAS frameworks struggle with reliability and scalability, especially on complex tasks. We present AgentGit, a framework that brings Git-like rollback and branching to MAS workflows. Built as an infrastructure layer on top of LangGraph, AgentGit supports state commit, revert, and branching, allowing agents to traverse, compare, and explore multiple trajectories efficiently. To evaluate AgentGit, we designed an experiment that optimizes target agents by selecting better prompts. We ran a multi-step A/B test against three baselines -- LangGraph, AutoGen, and Agno -- on a real-world task: retrieving and analyzing paper abstracts. Results show that AgentGit significantly reduces redundant computation, lowers runtime and token usage, and supports parallel exploration across multiple branches, enhancing both reliability and scalability in MAS development. This work offers a practical path to more robust MAS design and enables error recovery, safe exploration, iterative debugging, and A/B testing in collaborative AI systems.

Abstract--Driven by the unceasing development of maritime services, tasks of unmanned aerial vehicle (UA V)-assisted maritime data collection (MDC) are becoming increasingly diverse, complex and personalized. As a result, effective task allocation for MDC is becoming increasingly critical. In this work, integrating the concept of spatial crowdsourcing (SC), we develop an SCbased MDC network model and investigate the task allocation problem for UA V-assisted MDC. In variable maritime service scenarios, tasks are allocated to UA Vs based on the spatial and temporal requirements of the tasks, as well as the mobility of the UA Vs. T o address this problem, we design an SCbased task allocation algorithm for the MDC (SC-MDC-T A). The quality estimation is utilized to assess and regulate task execution quality by evaluating signal to interference plus noise ratio and the UA V energy consumption. The reverse auction is employed to potentially reduce the task waiting time as much as possible while ensuring timely completion. Additionally, we establish typical task allocation scenarios based on maritime service requirements indicated by electronic navigational charts. Simulation results demonstrate that the proposed SC-MDC-T A algorithm effectively allocates tasks for various MDC scenarios. Furthermore, compared to the benchmark, the SC-MDC-T A algorithm can also reduce the task completion time and lower the UA V energy consumption. RIVEN by the continuous development of maritime services such as marine resources exploration, reconnaissance and surveillance, anti-submarine, marine tourism, marine transportation and emergency collection, tasks of unmanned aerial vehicle (UA V)-assisted maritime data collection (MDC) are becoming increasingly diverse, complex and personalized [1]-[3]. Personal use of this material is permitted.