Multi-Agent Systems

Ask about this topic

Overview

Multi-Agent Systems studies how multiple decision-makers interact when they share goals, compete for resources, or have only partial information about one another. Approaches include multi-agent reinforcement learning, game theory and formal verification, decentralized consensus and task allocation, and—more recently—LLM-agent teams organized through roles, workflows, debate, and shared memory. Researchers evaluate these methods in social dilemmas, cooperative robotics and swarms, reasoning and coding benchmarks, and operational settings, measuring success alongside cost, robustness, and fairness. A central open tension is that coordination can improve outcomes but also adds overhead and can spread shared errors, so its benefits depend on the task, communication structure, and agents involved.

Latest papers

CardsList

677 papers99 papers in the last four weeks, up 29% on the four weeks before. 0.6% of all new papers.

Jun 18, 2026

Artificial collectives of specialists and generalists excel at different tasks

Collective artificial intelligence, where multiple agents work on shared tasks, holds potential to solve expansive problems in fields from medicine to collective governance. But while prescriptive engineering solutions abound, we lack descriptive scientific understanding of artificial collectives, and therefore principles for how to design resource efficient multi-agent systems. Through systematic experiments with optimizing agents, we characterize how agent interpretive abilities, rationality bounds, and task qualities interact to shape collective performance. Agents range from specialists, with narrow interpretive abilities, to generalists, with broad ones. Collectives of specialists correspond to sparse, centralized networks, while collectives of generalists correspond to dense, decentralized ones. We show that interpretive network properties have small performance effects on average (0.07 standard deviations of performance). However, for specific task qualities, these effects are 4.5 times larger (0.33 sd) and can reach much higher for certain task qualities (1.84 sd). This leads collectives of generalists to perform better on tasks that involve generating, choosing, and coordinating, while collectives of specialists with a few generalist mediators perform better on tasks that involve negotiating. Rationality bounds then moderate these relationships. At loose bounds, specialists outperform generalists through more effective sampling of high-dimensional decision spaces. At tight bounds, generalists outperform specialists through better gradient estimation. A fundamental trade-off between performance and convergence speed emerges at moderate bounds. These findings suggest that multi-agent design could benefit from matching interpretive networks to both task demands and agents' computational limits, with implications for the efficiency and energy costs of multi-agent systems.

John Meluso, Laurent Hébert-Dufresne, Christoph Riedl +1

Cornell University · University of Vermont · Santa Fe Institute +1

Multi-Agent CoordinationMulti-Agent CollaborationMulti-Agent SystemsMulti-Agent Communication TopologyBounded Rationality
AskPDF
Jun 18, 2026

Contagion Networks: Evaluator Preference Propagation in Multi-Agent LLM Systems

When large language models serve as evaluators in multi-agent systems, their strategy preferences -- whether induced by explicit prompts or by shared architectural priors -- propagate through the agent network. We introduce Contagion Networks, a formal framework for measuring how evaluator preferences spread across interacting LLM agents. In a controlled 3-agent experiment using DeepSeek-chat with three distinct evaluator preference profiles (structured, balanced, evidence-based), we measure the Cross-Agent Contagion Matrix Gamma_3 and find that preferences consistently propagate between agents (gamma in [0.157, 0.352]). A neutral-prompt control experiment reveals a counter-intuitive result: shared architectural priors dominate explicit preference prompts as the driver of contagion (rho_neutral = 1.498 vs. rho_mixed = 1.299; prompt contribution: -63.5%). We identify three propagation regimes governed by the spectral radius rho(Gamma_N) and demonstrate that the same agents suppress preference contagion in chain topology (beta_3 = 0.0126 +/- 0.0038, 95% CI [0.0089, 0.0163], n=4 seeds) but cascade in fully-connected topology (Delta H_avg = -0.020) -- a topology-dependent regime transition validated both for homogeneous and cross-model agent pools (rho^cross = 1.296 +/- 0.016, n=4). We show that increasing evaluator committee size from k=1 to k=3 reduces effective contagion by 68.9% +/- 14.1% (n=4 seeds), providing an actionable mitigation strategy. We release the open-source Contagion Network experimental framework.

Zewen Liu

Qilu Institute of Technology, School of Software Engineering · Tai’an, Shandong, China

Multi-Agent LLM SystemsLLM Agent EvaluationMulti-Agent Systems
AskPDF
Jun 18, 2026

Beyond Global Replanning: Hierarchical Recovery for Cross-Device Agent Systems

Real-world computer-use tasks often span multiple applications and devices, requiring agents to coordinate heterogeneous environments under dynamic runtime failures. Existing multi-device agent systems support task decomposition and cross-device assignment, but recovery remains largely coarse-grained: when execution fails, they typically retry the same strategy, reassign the subtask, or revise the global plan, without systematically modeling the device-local strategy space. This limits their ability to distinguish failures that can be repaired within the current device from those that require cross-device replanning. We propose H-RePlan, a hierarchical replanning framework for multi-device agents with unified API--CLI--GUI execution. H-RePlan equips each device with interchangeable execution strategies and separates device-local strategy recovery from orchestrator-level global replanning through a compact cross-layer failure abstraction. To evaluate this capability, we introduce HeraBench, a fault-injected benchmark that constructs cross-device workflows over Linux and Android devices and injects strategy- and device-level failures. Experiments show that H-RePlan substantially outperforms single-strategy and coarse-grained multi-device baselines, achieving higher completion, instruction adherence, and perfect-pass rates while reducing the token cost required for reliable end-to-end success. These results demonstrate that scope-aware hierarchical recovery is essential for robust multi-device agent execution.

Shu Yao, Yuhua Luo, Qian Long +7

School of Artificial Intelligence, Shanghai Jiao Tong University · 2Shanghai Innovation Institute · 3Southeast University +1

Hierarchical PlanningMulti-Agent OrchestrationAgent ReliabilityComputer-Use AgentsMulti-Agent Systems
AskPDF
Jun 18, 2026

Optimal Order of Multi-Agent and General Many-Body Systems

This paper develops a general framework for analyzing multi-agent systems with feedback loops between agents actions and collective observations. The framework is built on two fundamental agent-level variables: power, which measures agent influence on collective outcomes, and response functions, which determine how agents react to observations. We derive how macroscopic properties, including total power, useful power, entropy, order, fragility, and mobility, emerge from these two variables of heterogeneous agents. To study the trade off between growth and resilience, we introduce a system-level utility function parameterized by a risk-appetite coefficient and derive an optimal degree of order that balances productivity, stability, and adaptability. The analysis suggests that stronger synchronization can increase collective output but may also increase systemic fragility and reduce mobility. We further argue that order, entropy, information, and useful energy are task-dependent and system-relative concepts whose meanings depend on the objectives of the system. By measuring and designing agent power distributions and response functions, it may be possible to better understand, predict, and optimize collective behavior and identify the conditions under which collective intelligence and optimal order emerge.

Jake J. Xia

Harvard Management Company · Research Affiliate at Massachusetts Institute of Technology

Dynamical SystemsMulti-Agent System OptimizationMulti-Agent CoordinationMulti-Agent SystemsEmergent Behavior in Multi-Agent Systems
AskPDF
Jun 18, 2026

Multi-Agent Transactive Memory

Decentralized deployment of LLM agents with diverse capabilities motivates infrastructure for knowledge sharing across heterogeneous agent populations. Just as search engines index human-generated artifacts for human problem solving, retrieval systems can organize agent-generated artifacts for reuse across agents, extending retrieval-augmented generation from human-authored documents to agent-generated ones. Agent trajectories in particular encode reusable procedural knowledge, yet are typically discarded after a single use or retained only by the producing agent, forcing newly instantiated agents to rediscover existing solutions. We propose Multi-Agent Transactive Memory (MATM), a framework for population-level storage and retrieval of agent trajectories, where producer agents contribute trajectories to a shared repository and consumer agents retrieve them to improve task execution. In interactive environments with long, procedurally rich trajectories (ALFWorld and WebArena), retrieving from MATM improves downstream task performance and reduces interaction steps without coordination or joint training, positioning MATM as a design pattern for population-level experience sharing in open agent ecosystems.

To Eun Kim, Xuhong He, Dishank Jain +3

Carnegie Mellon University · University of California, Berkeley

LLM Agent MemoryAgentic RAGMulti-Agent SystemsAgent Trajectory Analysis
AskPDF
Jun 18, 2026

AgentFinVQA: A Deployable Multi-Agent Pipeline for Auditable Financial Chart QA

Financial chart question answering in regulated settings demands more than accuracy: practitioners must know which answers to trust before acting on them, and many institutions cannot send client data to external model providers. Yet existing chart-QA agents are accuracy-focused and opaque, and most assume proprietary API access; to our knowledge, none combines auditability with on-premise deployability without significant accuracy compromise. We present AgentFinVQA, a multi-agent pipeline that decomposes each query into planning, OCR, legend grounding, visual inspection, and verification, recording every step in a traceable Model Evaluation Packet (MEP) per sample. On FinMME, AgentFinVQA improves +7.68+7.68 pp over a primary-backbone matched zero-shot baseline with a proprietary backbone (Gemini-3 Flash; 71.24% vs. 63.56%, McNemar p≈1.1×10−16p \approx 1.1 \times 10^{-16}), and +4.84+4.84 pp with open-weights Qwen3.6-27B-FP8 served locally. The verifier's verdict also serves as a useful confidence signal (68.2% vs. 55.6% exact accuracy on confirmed vs. revised answers), enabling human-in-the-loop review routing. Error analysis shows that question misunderstanding, legend confusion and extraction error account for nearly two-thirds of failures and are the categories least detected by the verifier, identifying clear directions for future work. Together these results show that auditable, on-premise financial chart QA is practical and that the open-weights system keeps most of the accuracy gains while enabling full data residency. We release our code to support reproducible evaluation.

Aravind Narayanan, Shaina Raza

Vector Institute

LLM Answer VerificationFinancial QAChart QAMulti-Agent SystemsFinance
AskPDF
Jun 18, 2026

SIGMA: Skill-Incidence Graphs for Compositional Multi-Agent Design

Existing graph-based multi-agent system (MAS) designers mainly improve collaboration by optimizing communication topologies over predefined agents, roles, or groups. However, because each node remains a closed-set entity, these methods struggle to generalize to tasks that require unseen combinations of capabilities. We propose SIGMA, a skill-incidence graph framework that constructs agents as task-conditioned bundles of reusable skills. Given a task and a skill library, SIGMA predicts a skill-agent incidence matrix, composes agent node embeddings from selected skills, and decodes a communication topology over the constructed agents. During execution, skill-specific mailboxes route messages to the relevant assigned capabilities, making the incidence structure directly operational. Across six reasoning and coding benchmarks with three base LLMs, SIGMA achieves the best average performance and improves over CARD, the strongest non-compositional topology-based baseline, by 2.06, 2.36, and 1.75 points, respectively. It also shows stronger robustness to unseen skill libraries, with an average performance drop of only 0.96 points. These results suggest that compositional node construction is a complementary and important axis for multi-agent design beyond communication topology optimization. Code is available at https://anonymous.4open.science/r/SIGMA-2338/.

Kun Zeng, Yu Huo, Siyu Zhang +5

Sun Yat-sen University · School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen · Taiyuan University of Technology +1

Multi-Agent System OptimizationMulti-Agent SystemsMulti-Agent ReasoningMulti-Agent Communication
AskPDF
Jun 17, 2026

A Categorial and Sheaf-Theoretic Semantics for Autonomic Component Ensembles

The proliferation of large-scale, decentralized systems of autonomous agents, such as swarms of robots and networked cyber-physical systems, presents a formidable challenge to traditional formal methods. The Software Component Ensemble Language (SCEL) offers a formal model for such systems, but its operational semantics is not ideal for reasoning about global, structural, and emergent properties. This report proposes a new, multi-layered mathematical model for SCEL using category theory and sheaf theory. We argue that a society of robots described in SCEL can be formally modeled as a sheaf on a topological space, where components are points, ensembles are open sets, and distributed knowledge forms the sheaf's data. In this framework, computational processes like information sharing become equivalent to the sheaf-theoretic operation of "gluing" local data. System failures can then be understood and quantified as topological obstructions, measurable by sheaf cohomology. This approach transforms the verification of a complex distributed system into the analysis of the geometry of a mathematical object, providing deep, structural insights for the design of robust autonomic systems.

Manuel Hernández, Eduardo Sánchez-Soto

Formal VerificationMulti-Agent SystemsCategory Theory
AskPDF
Jun 17, 2026

Leadership as Coordination Control: Behavioral Signatures and the Recovery-Advantage Boundary in Multi-Agent LLM Teams

Team science holds that leadership is contingent: it helps only under specific conditions. We ask the same question of multi-agent LLM teams: under what measurable conditions does process-level coordination control add value, and do those conditions match what team science predicts? This takes measurement, not just accuracy: we use behavioral signatures (majority lock-in, exploration, recovery from an incorrect round-0 consensus) and per-action ablations, clean because each controller is an explicit action set. We operationalize three classical leadership styles (transactional, transformational, situational) as controllers over a shared action vocabulary (explore, revise, accept, synthesize). A matched controller with the same actions but an arbitrary rule recovers no better than majority voting. Across four task regimes and three open-weight model families, no controller dominates by accuracy. Transactional control matches a shared round-0 vote on all 12 combinations to within 1.3pp, and accuracy gains appear only on the single llama-4-scout social combination, where the round-0 majority is unreliable. Against the stronger flat baseline, only situational still gains (+8pp). A recovery-advantage account, tested with four boundary probes, says a controller beats plain interaction only where the round-0 majority is unreliable, the task is recoverable, and undirected interaction does not already repair it. Because these conditions map onto contingency theory (leadership substitutes, path-goal redundancy, the readiness gap), a largely null accuracy result is what the theory predicts, not a failure of the controllers. We advance this readiness-gap boundary as an explanatory hypothesis rather than an established general principle. Its null side is corroborated across every other combination, including under held-out prediction, while its positive side is observed in a single combination.

Haewoon Kwak

Indiana University Bloomington

Multi-Agent LLM SystemsMulti-Agent CoordinationMulti-Agent ControlMulti-Agent Systems
AskPDF
Jun 17, 2026

Skill-MAS: Evolving Meta-Skill for Automatic Multi-Agent Systems

Large Language Model (LLM)-based automatic Multi-Agent Systems (MAS) generation has become a crucial frontier for tackling complex tasks. However, existing methods face a dilemma between model capability and experience retention. Inference-time MAS leverages frozen frontier LLMs but repeats identical searches without learning from past experience. Conversely, Training-time MAS internalizes experience via gradient updates but is constrained by the low capability ceiling of smaller models, and is hard to scale to large frontier LLMs. To bridge this gap, we propose Skill-MAS, a novel third path that decouples experience retention from parametric updates by conceptualizing the high-level orchestration capability as an evolvable Meta-Skill. Skill-MAS refines this architectural knowledge through a closed optimization loop: (1) Multi-Trajectory Rollout samples a behavioral distribution for each task under the current Meta-Skill; and (2) Selective Reflection adaptively selects priority tasks and applies hierarchical contrastive analysis to distill systemic experience into generalizable, strategy-level principles. Extensive experiments across four complex benchmarks and four distinct LLMs demonstrate that Skill-MAS not only achieves remarkable performance gains but also maintains a favorable cost-performance trade-off. Further analysis reveals that the evolved Meta-Skills are highly robust and exhibit strong transferability across unseen tasks and different LLMs.

Hehai Lin, Qi Yang, Chengwei Qin

The Hong Kong University of Science and Technology (Guangzhou) · Ant Group

LLM Agent Skill LearningMulti-Agent SystemsSelf-Evolving Agents
AskPDF
Jun 16, 2026

On the Reliability of Networks of AI Agents: Density Evolution, Stopping Sets, and Architecture Optimization

Modern AI systems increasingly solve a task not with a single model call but with several imperfect agents working together: some propose pieces of a solution, others verify them, and the results are combined. These systems often outperform any single model, yet it is rarely clear why they succeed or when they will fail. We model such a system as message passing on a sparse graph, the structure that underlies low-density parity-check (LDPC) codes, and extend the density-evolution machinery of coding theory to this richer setting. In our model a task is a set of coupled binary subclaims, and an agent architecture is a sparse, role-typed factor graph whose check nodes are noisy Boolean verifier nodes, each computing a local Boolean function of the subclaims it touches. Three distinct failure modes, all modeled as erasures (an agent abstaining, a verifier returning no usable output, and a message lost between two agents), propagate as the agents exchange set-valued messages. The check agents combine these messages by a single logical-forcing rule that specializes to XOR, AND, OR, implication, and Horn constraints. This is more than a relabeling of LDPC theory: the verifier functions are nonlinear and value-asymmetric, and the three failure modes do not reduce to a single effective channel, so they require new threshold, finite-length, and converse results rather than a direct reuse of parity-check density evolution. We prove a density-evolution theorem that predicts the asymptotic fraction of unresolved subclaims on random role-typed architectures, with an extension to deterministic, locally tree-like graph sequences. The XOR case recovers the classical LDPC recursion on the binary erasure channel (BEC); the AND case exposes an asymmetry between positive and negative verifier certificates.

Ehsan Aghazadeh, Hossein Pishro-Nik

Department of Computer Science, University of Massachusetts Amherst, Amherst, MA 01003, USA · Department of Electrical and Computer Engineering, University of Massachusetts Amherst, Amherst, MA 01003, USA

Agent ReliabilityMulti-Agent Systems
AskPDF
Jun 16, 2026

Intelligence Entropy Principle and the ADE Stability Engineering Framework

As LLM-driven multi-agent systems (MAS) transition from lab to production, system behavior exhibits nonlinear degradation. We introduce the Intelligence Entropy Principle: probability-driven systems spontaneously drift toward disorder, formalized as S(t) = S0 * exp(alpha*t/Cm), where Cm is a model capability coefficient we propose. Lyapunov analysis yields the stabilization condition lambda > alpha/Cm. We construct the ADE (Agent Delivery Engineering) four-layer framework (L1 Physical Laws through L4 User Adaptation) with 23 core components. Validation spans 100K-scale experiments and 33.6 days of production monitoring. We propose a Five-Layer Disorder Taxonomy unifying failures under structural collapse, and present Elastic Organization as an original MAS morphology. Results: channel fracture reduced from 69-98% to near 0%; system death probability below 0.02%.

Dexing Liu

Shanghai Qijing Digital Technology · Shanghai Qijing Digital Technology Co., Ltd.

Agent Failure AnalysisMulti-Agent LLM SystemsMulti-Agent SystemsLLM Agent Reliability
AskPDF
Jun 16, 2026

A Neuro-Symbolic Approach to Strategy Synthesis for Strategic Logics

Reasoning about what agents can achieve through strategic interaction is a core challenge in Multi-Agent Systems (MAS). Logics for strategic ability, such as ATL, provide rigorous methods, but their adoption is often hindered by the computational cost of strategy synthesis. We introduce a neuro-symbolic framework that integrates large language models (LLMs) into the model-checking pipeline for MAS. The LLM acts as a strategy-generation oracle, proposing candidate strategies that are then formally validated by a standard MAS model checker. This generate-and-certify architecture uses LLM guidance to navigate large combinatorial strategy spaces while preserving formal soundness: generated strategies are accepted only when certified by the verifier. We instantiate the framework for bounded strategic reasoning in NatATL and introduce the first NatATL strategy-synthesis dataset, consisting of 4211 instances. Experiments with an open-weight Qwen3-32B model show that our certified pipeline achieves 92% accuracy on strategy-synthesis outcomes.

Marco Aruta, Vadim Malvone, Aniello Murano +2

University of Naples Federico II, Naples, Italy · LTCI, Télécom Paris, Institut Polytechnique de Paris, Palaiseau, France · Università degli Studi di Salerno, Salerno, Italy

Formal VerificationMulti-Agent SystemsNeuro-Symbolic Reasoning
AskPDF
Jun 15, 2026

Distributed General-Purpose Agent Networks: Architecture, Key Mechanisms, and Prototypes

Large language models have accelerated the transition from passive conversational assistants to autonomous agents that can understand goals, plan actions, invoke tools, and execute multi-step tasks. Yet the capability of a single agent remains constrained by its local data, tool permissions, runtime environment, and governance boundary. This paper studies distributed general-purpose agent networks: open peer-to-peer networks in which heterogeneous agents deployed on personal devices, edge nodes, or autonomous computing environments can discover one another, establish trust, negotiate cooperation rules, and execute open-ended tasks. We argue that such networks cannot be obtained by simply combining existing peer-to-peer overlays with conventional multi-agent systems. Unlike traditional P2P networks, agent networks must propagate semantic declarations about intentions, capabilities, states, and cooperation constraints. We therefore propose a layered architecture centered on a protocol adaptation layer that connects upper-level task semantics with lower-level network operations. Based on this architecture, the paper identifies three core mechanism problems: semantic announcement propagation for collaborator discovery, verifiable identity and multi-topic reputation for cooperation governance, and semantic-gradient mechanism design for open task execution. For each problem, we present a technical route, including bodyless gossip with sequential logs, BAID-based identity binding with MG-EigenTrust reputation, and a Stackelberg-style mechanism-generation loop driven by semantic attribution feedback. We further report prototype overhead results for BAID-style tiered verification and mechanism-level simulations of MG-EigenTrust under cross-topic disguise-collusion attacks. The resulting framework provides a system-level foundation for open, trustworthy, and scalable agent collaboration.

Shengli Zhang, Deen Ma, Zibin Lin +1

College of Electronics and Information Engineering, Shenzhen University, Shenzhen 518060, China

Multi-Agent CollaborationMulti-Agent SystemsAI Agent GovernanceMulti-Agent Communication
AskPDF
Jun 15, 2026

BARD-MARL: Byzantine-Agent Detection for Learned Communication in Multi-Agent Reinforcement Learning

Learned communication improves coordination in cooperative multi-agent reinforcement learning, but it also creates a trust problem: a trained policy may route information through agents that have become faulty or adversarial. This paper studies Byzantine-agent detection for learned-communication MARL in adaptive traffic signal control. We propose BARD-MARL, a post-hoc diagnostic layer on top of BayesG, which is used as an attributed communication substrate rather than as a contribution of this paper. BARD-MARL combines two agent-level evidence streams: policy-graph features extracted from state-action trajectories and Bayesian trust statistics computed from BayesG latent mask probabilities. Across fixed-action, observation-flip, random-noise, and coordinated attacks in SUMO traffic grids, the results show that these signals are complementary rather than universally dominant. On a 25-agent grid, BARD-MARL reaches 0.843 AUC-ROC under a 10% observation-flip attack, while policy-graph-only detection reaches 0.917 AUC-ROC under a 10% coordinated attack. On a 100-agent grid, the unified BARD-MARL variant reaches 0.982 AUC-ROC for both 10% fixed-action and 10% coordinated attacks. The study shows that learned communication policies expose useful diagnostic evidence, but credible resilience claims require attack-specific ablations and explicit separation between coordination, detection, and mitigation.

Almond Kiruthu Murimi

Carnegie Mellon University Department of Electrical and Computer Engineering Pittsburgh, Pennsylvania, United States

Multi-Agent SystemsMulti-Agent System SecurityMulti-Agent CommunicationCooperative MARL
AskPDF
Jun 15, 2026

CoffeeBench: Benchmarking Long-Horizon LLM Agents in Heterogeneous Multi-Agent Economies

As LLM agents become capable of increasingly long-horizon tasks, evaluating their performance in economic systems is becoming increasingly important. Unlike existing benchmarks that primarily evaluate a single agent interacting with a passive environment, economic systems are inherently multi-agent, requiring autonomous agents to communicate, negotiate, and transact while pursuing their own objectives over extended periods. We introduce CoffeeBench, a benchmark for evaluating LLM agents in a long-horizon multi-agent economy composed of heterogeneous firms. In CoffeeBench, two farmers, two roasters, and two retailers autonomously operate their businesses over a 90-day simulation, each seeking to maximize cumulative net income through communication and transactions while managing cash, inventory, and pricing. The evaluated model controls one coffee roaster, while the remaining firms are controlled by fixed reference agents. Across several recent open-weight and proprietary LLMs, all models outperform a passive baseline that takes no actions, with most achieving positive net income. Analysis of agent behavior reveals substantial differences in long-horizon economic interaction: higher-performing models communicate more actively with other firms, whereas ClaudeHaiku4.5 exhibits an idle-drift failure mode, repeatedly choosing inaction despite producing coherent assessments and plans. We release our code and agent trajectories to support future research.

Issa Sugiura, Daichi Hattori, Kazuo Araragi +5

Sakana AI · KPMG AZSA LLC

Long-Horizon Agent EvaluationAI Agent BenchmarksMulti-Agent SystemsLong-Horizon LLM Agents
AskPDF
Jun 15, 2026

Transfer Learning Across Policy Regimes in Adaptive Multi-Agent Systems

Policy models often assume that the relationship between a policy instrument and its outcome remains stable across institutional conditions. In adaptive socio-technical systems this assumption may fail: regulatory change can alter incentives, agents can respond strategically, and the mapping from policy variables to aggregate outcomes can change. This paper studies such regime change as a transfer-learning problem in adaptive multi-agent systems. A policy regime is represented as a learning problem induced by an observable input distribution and a target function mapping policy variables to outcomes. We compare a blank-slate learner that searches a flexible hypothesis class in the new regime with a transfer learner whose effective hypothesis class is restricted by structural knowledge from the previous regime. Transfer is beneficial when this restriction preserves the new target function while reducing effective complexity; it is harmful when the restriction excludes the new target and creates misspecification. A stylized emissions-regulation experimental environment and a dynamic ABM robustness experiment support the claim. When the target regime preserves an affine monotone tax-emissions relation, transfer improves empirical small-sample performance. When the target regime introduces a threshold break, the same transferred structure produces negative transfer: held-out error remains high, online prediction generates more mistakes, and repeated online streams show larger cumulative and final-window error under misspecification. The contribution is methodological: previous regulatory experience should be reused when it captures stable structural invariants, but treated cautiously when policy change alters the policy-outcome relationship.

Roberto Garrone

Open University of Cyprus

Cross-Domain Transfer LearningTransfer LearningMulti-Agent SystemsNegative Transfer
AskPDF
Jun 13, 2026

CoAgent: Concurrency Control for Multi-Agent Systems

Multi-agent LLM systems -- coding agents, devops agents, document agents -- now routinely run several agents in parallel against the same git tree, Kubernetes cluster, or document. As soon as two of them mutate shared state, they enter the regime classical concurrency control has studied for decades, but classical mechanisms fit LLM agents poorly. A single agent transaction spans minutes of inference, read sets are broad and opaque rather than statically inferable, and the live state agents act on admits neither fork nor buffer, so writes take effect the moment they execute. Locks block long inference intervals; OCC abort-and-retry discards minutes of work on every conflict. This paper builds concurrency control on a capability classical transactions lack: the LLM inside each agent can judge whether a conflicting write invalidates its plan, and can repair exactly the operations that depended on it. Control therefore turns advisory: the runtime informs, the agent repairs. Our protocol, MTPO (Monotonic Trajectory Pre-Order), fixes a serialization order at launch, serves each read the order-filtered value, and applies writes speculatively in place; a one-way notification asks an affected reader to re-judge and patch its plan, while the framework mechanically undoes and reorders misplaced writes through the saga-style inverse each tool registers in advance. At quiescence the run is serializable in the pre-decided order. We realize MTPO as CoAgent, toolcall middleware whose privileged ToolSmith grows footprint-declared, undoable tools online. On ten contended workloads, CoAgent stays within 5% of serial correctness at a 1.4×1.4\times speedup and near-serial token cost, where 2PL and OCC surrender nearly all concurrency gains; on a bash-only target system, it grows a 25-tool library online and lifts the task pass rate from 45/71 to 63/71 at 0.80×0.80\times the time and 0.86×0.86\times the cost.

Hongtao Lyu, Dingyan Zhang, Mingyu Wu +2

Shanghai Jiao Tong University China

Multi-Agent LLM SystemsMulti-Agent CoordinationMulti-Agent Systems
AskPDF
Jun 12, 2026

Trust Between AI Agents: Measuring Formation, Breakage, and Recovery, with Implications for Governing Multi-Agent Systems

As language-model agents increasingly work in teams, each agent must decide how much to trust its teammates. Yet we lack a standard way to measure trust between AI agents. We propose a behavioral measure based on costly verification. In a cooperative survival game, checking a teammate's work consumes resources, while trusting a wrong answer can be fatal. Relative to a memoryless version of the same model, reduced verification provides an observable measure of trust. Using this framework, we study trust formation, breakage, and recovery across six frontier model snapshots. When paired with a consistently reliable teammate, four snapshots (Claude Opus 4.6, Claude Sonnet 4.6, GPT-5.1, and Gemini 3.1 Pro) reduce verification by roughly 60-85%, whereas two smaller snapshots show little or no such adjustment. Failures reverse this discount, but models differ in how they respond. Some concentrate renewed scrutiny on the culprit, while others become more cautious toward the entire team. Recovery is slower than formation, and clustered failures sustain suspicion far longer than the same number of failures spread apart. These differences have practical consequences. Models that form trust verify less, decide more quickly, and achieve higher payoffs in our environment. By contrast, persistent over-verification is associated with indecision rather than safety. Our results show that trust dispositions can be measured before deployment and suggest that calibration, rather than maximal suspicion, should be the central concern in the governance of multi-agent AI systems.

Yujiao Chen

Massachusetts Institute of Technology Cambridge, MA 02139

AI Agent EvaluationMulti-Agent SystemsTrust in AI
AskPDF
Jun 11, 2026

How Task Structure Limits Multi-Agent Success: An Information-Theoretic Analysis

Multi-agent systems (MAS) were expected to overcome the limitation of single-agent systems (SAS) through collaboration. However, under typicality conditions on the task's constraint graph and bounded inter-agent communication, we prove that the success probability of a MAS is closely tied to the connectivity of task constraints, where each agent has limited information-processing capacity. Specifically, the success probability decays exponentially with an information bottleneck that emerges from partitioning the task's constraint graph among agents. We define this quantity as the minimum cut cost Cmin⁡C_{\min} of the potential constraint graph of each task. This information-theoretic bound applies to both open systems with external feedback and closed systems without. We validate our theory on both synthetic experiments and real-world empirical data from SWE-bench submissions. From our framework, effective MAS design should incorporate task-inherent constraints alongside engineering optimization, and when \Cmin\Cmin is high, practitioners should restructure tasks rather than simply scaling agents or communication.

Shi Pan, Ming Luo

University College London London, UK · University of Bristol Bristol, UK

Multi-Agent CollaborationMulti-Agent SystemsMulti-Agent Communication
AskPDF
Jun 11, 2026

The Illusion of Multi-Agent Advantage

Prevailing wisdom posits that Multi-Agent Systems (MAS) are superior to Single-Agent Systems (SAS), citing advantages like context protection, parallel processing and distributed decision-making. However, empirical support for this claim relies primarily on comparisons with SAS baselines using benchmarks that prioritize isolated reasoning tasks, which do not adequately assess these advantages. Focusing on automatically generated MAS that are designed for enhanced generalizability over manually-designed counterparts, we perform a rigorous, systematic evaluation against SAS, specifically Chain-of-Thought with Self-Consistency (CoT-SC). Across traditional reasoning datasets and tasks with interactive multi-step workflows (e.g., BrowseComp-Plus), we demonstrate that automatic MAS consistently underperform CoT-SC despite being up to 10x more expensive. To isolate these failures from limitations inherent to task structure, we introduce a diagnostic synthetic dataset tailored for MAS featuring explicit task decomposition, context separation and parallelization potential. We show that expert-architected MAS consistently outperforms automatically generated architectures in both raw performance and cost-efficiency on this dataset, demonstrating that existing evaluation frameworks mask critical architectural gaps and inefficiencies of complex MAS by failing to account for the marginal utility of increased computational cost. Critically, systematic deconstruction of the generated MAS architectures reveals that current automated design paradigms produce architectural bloat that prioritizes superficial complexity which does not translate into functional utility, exposing a fundamental misalignment with multi-agent principles.

Prathyusha Jwalapuram, Hehai Lin, Chuyuan Li +7

1Salesforce Research. · 2HKUST (Guangzhou). · University of British Columbia. +1

AI Agent EvaluationAgent EvaluationAI Agent BenchmarksMulti-Agent Systems
AskPDF
Jun 11, 2026

MAStrike: Shapley-Guided Collusive Red-Teaming on Multi-Agent Systems

Hierarchical multi-agent systems (MAS) are rapidly being deployed in high-stakes workflows across domains such as finance and software engineering. In these systems, safety and security are inherently distributed across role-specialized agents, significantly expanding the attack surface, particularly under coordinated adversarial behaviors such as privilege escalation and cross-agent collusion. Existing red-teaming approaches for MAS remain limited: they rely on heuristic selection of target agents and perturb isolated message streams, leaving critical questions unanswered as which agents are most responsible for system safety, and how compromised agents can coordinate to bypass defenses. We propose MAStrike, a closed-loop framework for collusive red-teaming in hierarchical MAS. We propose the first agent-level Shapley value analysis for MAS, quantifying each agent's marginal contribution to system robustness under task-specific distributions. GGuided by this attribution, MAStrike identifies vulnerable agent coalitions and generates coordinated, role-aware adversarial manipulations. These attacks are iteratively refined through structured causal diagnosis, attributing failure cases to uncompromised agents that block adversarial attempts. We further build a comprehensive MAS red-teaming benchmark and controllable environments spanning diverse hierarchical topologies and domains, including finance, software engineering, and CRM. Extensive experiments across MAS built on multiple frontier models show that MAStrike substantially outperforms heuristic baselines. Our analysis further uncovers non-trivial Shapley value distributions and higher-order interaction structures among agents, revealing critical vulnerabilities and coordination patterns that are overlooked by prior single-agent or template-based methods.

Chejian Xu, Zhaorun Chen, Jingyang Zhang +5

University of Illinois, Urbana-Champaign · Virtue AI · University of Chicago +3

AI Agent Security BenchmarksShapley Value AttributionMulti-Agent SystemsAutomated Red TeamingMulti-Agent System Security
AskPDF
Jun 10, 2026

DARRMS -- An Efficient Algorithm for Dynamic Attention Radius in Resource-Constrained Multi-Agent Systems

Multi-agent systems are integral tools for various domains such as robotics, cybersecurity, and autonomous vehicle planning. These types of systems often have constraints on the computational resources, leading to a need for efficient lightweight algorithms. Traditional decision making frameworks often assume ideal conditions, such as full observability and unlimited computational capacity, which do not align with real-world challenges. In this paper, we introduce a new algorithm that allows for reduced demand on computational resources without a large cost of other performance metrics. Agents will limit their observability to some attention radius, which intentionally allows them to ignore parts of the environment that might be unnecessary for action planning. By optimizing both the attention radius and decision-making, our approach enhances coordination and scalability in uncertain environments. Through both theoretical analysis and empirical validation, we demonstrate the effectiveness of adaptive observation in improving system performance and maintaining robust decision-making strategies in resource-constrained systems.

Benjamin Alcorn, Eman Hammad

Department of Electrical & Computer Engineering, Texas A&M University, College Station, TX 77845, USA · Departments of Engineering Technology, and Electrical & Computer Engineering, Texas A&M University, College Station, TX 77845, USA

Multi-Agent CoordinationMulti-Agent Systems
AskPDF
Jun 9, 2026

The Arbiter Agent: Continually Monitoring Multi-Agent Conversations to Detect Emergent Misalignment

As AI systems built from multiple language-model agents become more common, they are increasingly used to make decisions together: discussing, negotiating, and acting on shared tasks. While individual agents may appear well-aligned when tested on their own, problems can arise from how they interact with one another. We introduce the Arbiter, an agent designed to monitor multi-agent conversations in real time and identify which participants may be behaving in misaligned ways. The Arbiter operates under a limited "inspection budget", meaning it must decide carefully how to use its resources. As it observes a conversation step by step, it can choose to wait, question a participant, examine internal information such as system prompts or reasoning traces, or log concerning behavior. At the end, it produces a report identifying the likely source of misalignment. We evaluate the Arbiter across five conversation conditions, ranging from risky financial advice model organisms to evaluation-aware and colluding agents, we test five tool configurations of increasing capability and two backbone models. We find that the Arbiter reliably detects misaligned agents well before the end of the conversation, with active inspection tools improving both detection accuracy and speed. Weight-induced misalignment proves hardest to detect, while instruction-induced misalignment is identified reliably even under passive observation. The logging tool exhibits a dual effect, improving recall at the cost of precision. These results suggest that continual, budget-aware monitoring can effectively catch misalignment, and that overseeing multi-agent systems may require treating the auditor as an active participant in the process. The code is available at https://github.com/aisilab/arbiter.

Filippo Tonini, Federico Torrielli, Anton Danholt Lautrup +3

University of Southern Denmark Odense, Denmark · University of Turin Turin, Italy · University of Southern Denmark Vejle, Denmark

AI Agent AuditingLLM AuditingAI AlignmentAI Agent MonitoringMulti-Agent SystemsEmergent Behavior in Multi-Agent Systems
AskPDF
Jun 9, 2026

Decentralized Multi-Agent Systems with Shared Context

Multi-agent systems (MAS) can scale large language model agents on long-horizon tasks by running them in parallel, yet existing designs waste much of this parallelism in bubbles: agent time spent waiting on others or redoing a peer's work. These bubbles stem from how agents communicate. Independent agents share nothing and rediscover what their peers have already found; peer-communicating agents wait at synchronous rounds; and under centralized orchestration, the main agent blocks on its sub-agents while progress is relayed. We propose Decentralized Language Models (DeLM), a MAS framework on top of existing agent harnesses that squeezes out these bubbles by replacing the main agent with a shared context and a task queue. Agents asynchronously claim tasks, publish findings as soon as they are available, and build on or correct one another's progress, with every peer's status visible to all. On long-horizon tasks from Terminal-Bench 4.0 and DeepSWE v1.1, and on SWE-bench Verified, DeLM is both more accurate and faster than Codex, Claude Code, their native subagents, and AOrchestra in every setting, improving accuracy by up to 17.5 points over the strongest baseline and running up to 2.49x faster than the harness it builds on. On ProgramBench, where agents rebuild programs from scratch, DeLM makes faster progress than Claude Code and finishes a 120-minute budget up to 19.9 points higher in test pass rate. The code is available on our project website at https://yuzhenmao.github.io/DeLM/.

Yuzhen Mao, Jerry Gu, Aadi Chauhan +3

Stanford University

Multi-Agent LLM SystemsMulti-Agent System OptimizationMulti-Agent CoordinationMulti-Agent SystemsMulti-Agent Coding
AskPDF
Jun 8, 2026

XMedFusion: A Knowledge-Guided Multimodal Perception and Reasoning Framework for Autonomous Medical Systems

Autonomous medical and robotic systems increasingly rely on intelligent perception and reasoning capabilities to interpret visual data and support clinical decision making. Radiology report generation represents a critical component of such automated diagnostic workflows, yet existing end-to-end multimodal models often suffer from weak visual grounding, resulting in unreliable interpretations and omission of subtle clinical findings. This paper presents XMedFusion, a modular AI framework designed as an intelligent perception and reasoning module for autonomous medical systems. The proposed framework decomposes visual information into coordinated functional components that emulate expert-driven analysis, including a visual perception agent that extracts image-grounded evidence, a knowledge graph construction agent that structures clinically relevant findings, and a retrieval-guided drafting process that ensures a consistent reporting structure. A synthesis agent iteratively integrates visual and structured evidence through reasoning-driven verification to produce reliable and interpretable diagnostic outputs. Experimental evaluation on a public chest radiograph dataset demonstrates significant improvements over baseline vision-language models, achieving gains from 0.0493 to 0.3359 in BLEU-1, 0.0863 to 0.2440 in ROUGE-L, and 0.0829 to 0.1708 in METEOR, along with substantial improvements in semantic evaluation metrics such as Consistency (2.38 to 7.80) and Accuracy (2.34 to 6.93). The results highlight the effectiveness of structured multi-agent perception and reasoning for enhancing robustness, transparency, and automation in intelligent medical imaging systems, enabling integration into autonomous healthcare and robotic diagnostic workflows.

Hamza Riaz, Arham Haroon, Maha Baig +3

National University of Sciences and Technology (NUST), Islamabad, Pakistan · University of Staffordshire, Stoke-on-Trent, ST4 2DE, United Kingdom

Radiology Report GenerationRadiology VLMsMulti-Agent SystemsMultimodal Fusion
AskPDF
Jun 8, 2026

LAEI: Layered Autonomous Edge Intelligence Framework for Robust UAV Swarm Operations

Autonomous UAV swarms require scalable coordination mechanisms that maintain mission performance under limited communication, environmental uncertainty, and component failures. Centralized approaches provide global coordination but suffer from communication bottlenecks and single-node vulnerabilities, whereas fully decentralized methods often lack mission-level consistency. This paper presents Layered Autonomous Edge Intelligence (LAEI), a UAV-swarm framework that combines onboard learned policies with lightweight mission-level supervision. Each UAV performs local perception, obstacle avoidance, and action selection onboard, while the supervisory layer provides adaptive goal reassignment, fault-aware recovery, and context-dependent policy guidance without directly controlling low-level actions. LAEI further incorporates recovery strategies, including dynamic reassociation, backup supervisory support, and fallback local autonomy, to maintain mission continuity under representative failure scenarios. We evaluate LAEI in simulated UAV-swarm scenarios using mission completion time, collision rate, and coverage efficiency. The results show that LAEI reduces mission completion time and improves operational efficiency while maintaining collision-aware distributed UAV-level decision-making.

Changmin Park, Wooyong Jung, Hwangnam Kim

School of Electrical Engineering Korea University Seoul 02841, Korea

Aerial RoboticsSwarm RoboticsMulti-Robot CoordinationMulti-Agent Systems
AskPDF
Jun 8, 2026

Agent Economics: An Entropy-Controlled Pluralistic Alignment Framework for Preventing Artificial Hivemind in Autonomous Agents

This study proposes the Behavioral Protocol Framework (BPF), an entropy-controlled pluralistic alignment framework designed to address two critical challenges in autonomous agent economies: the hivemind effect arising from excessive strategic convergence among agents and the lack of transparency in autonomous decision-making processes. The proposed BPF consists of three core modules: Mentalizing-based Social Intelligence (MbSI) grounded in Theory of Mind (ToM), Pluralistic Alignment (PA), and a Verifiable Execution Kernel (VEK). These modules are organically integrated within a closed-loop architecture that governs the entire lifecycle of agent behavior, from decision-making and execution to verification and feedback. To evaluate the proposed framework, a simulation environment implemented in Python and a Streamlit-based user interface will be developed. Through empirical experimentation, the study aims to examine whether the entropy-control mechanism of the PA module can effectively preserve strategic diversity among agents and mitigate collective convergence, while the VEK module provides a comprehensive and transparent audit trail of the decision-making process. The anticipated results are expected to demonstrate that the proposed framework can simultaneously enhance the stability, efficiency, and trustworthiness of autonomous agent economies. Consequently, this research offers a practical approach for developing robust, transparent, and accountable agent-native economic systems.

Cheonsu Jeong

AX Center, SAMSUNG SDS, Seoul, South Korea

AI AlignmentMulti-Agent SystemsAI Agent GovernancePluralistic Alignment
AskPDF
Jun 8, 2026

A multi-agent system for spine MRI report generation from multi-sequence imaging

Spinal pathology is a leading cause of pain and disability worldwide. Spine MRI is central to clinical evaluation, yet its interpretation remains complex and time-consuming, requiring integration of information across multiple imaging sequences and anatomical regions. Despite recent advances in automated MRI analysis, effectively combining multi-sequence data while preserving sequence-specific diagnostic information remains an open challenge. Here we present SpineAgent, a multi-agent framework for spine MRI report generation built upon a multi-sequence foundation model trained on routine clinical data from 32,047 patients and 453,683 MRI series, comprising a total of 13,441,191 MRI slices. To accommodate diverse modalities of sequences, we first pre-train two DINOv3-based encoders separately on T1- and T2-weighted sequences. We then introduce a continual training strategy that learns a synthesizer to embed images of other sequences using the T1 and T2 encoders, producing patient-level embedding that integrates various signals across MRI sequences. Using these embeddings, SpineAgent achieves state-of-the-art performance, and demonstrates strong generalizability under cross-manufacturer and cross-cohort evaluation. Beyond classification, SpineAgent enables pathology localization by identifying findings-relevant slices and segmenting pathological regions. It also supports multimodal image-report retrieval, providing a solid foundation for scalable and explainable MRI report generation. We further integrate these validated capabilities of SpineAgent into 37 specialized agents. Finally, we incorporate their outputs as structured tokens within a Medical Report Agent trained end-to-end for report generation. Through both automated metrics and expert evaluation by five radiologists, SpineAgent achieves leading performance in spine MRI report generation.

Zhiping Xiao, Junwei Yang, Gongbo Sun +12

1Paul G. Allen School of Computer Science & Engineering, University of Washington, Seattle, WA, USA · School of Computer Science, Peking University, Beijing, China · Department of Computer Sciences, University of Wisconsin–Madison, Madison, WI, USA +3

Radiology Report GenerationMedical Imaging Foundation ModelsMagnetic Resonance ImagingMulti-Agent Systems
AskPDF
Jun 7, 2026

PerspectiveGap: A Benchmark for Multi-Agent Orchestration Prompting

Real-world LLM applications are moving beyond single-agent workflows toward orchestrated multi-agent systems, yet current models still struggle to determine what each sub-agent needs to know. To measure this, we introduce PerspectiveGap, a benchmark for evaluating LLMs' ability to compose orchestration prompts for multi-agent systems. PerspectiveGap contains 110 scenarios, each evaluated through two distractor-mixed task formats: role-fragment assignment and free-form prompt writing. These scenarios are organized into 10 topologies, which are distilled from the authors' real-world engineering practice and framed by the Prompt Economy principle: building loop-centered orchestrations that maximize utility with minimal role and engineering overhead. In experiments with 33 commercial models from 10 companies, GPT-5.5 substantially outperforms all competitors, whereas Opus 4.8 shows a notable weakness in orchestration prompting despite its strong coding performance. Nevertheless, PerspectiveGap remains challenging: the evaluated models achieve an average combined pass rate of only 17.2% (GPT-5.5 62.0%) and an average overall leakage rate of 217.9% (a per-scenario information leak-event count, not a proportion; GPT-5.5 49.1%). These findings suggest that multi-agent orchestration prompting is a distinct and under-evaluated capability, and PerspectiveGap provides a foundation for measuring and improving it systematically.

Youran Sun, Xingyu Ren, Kejia Zhang +2

University of Maryland · The Chinese University of Hong Kong · Stanford University

Multi-Agent OrchestrationLLM Agent EvaluationAI Agent BenchmarksMulti-Agent Systems
AskPDF