Multi-Agent LLM Systems

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

85 papers in the last four weeks, up 107% on the four weeks before. 0.8% of all new papers.

Jul 13Week of Sep 28

Latest papers 665

Apr 19, 2026cs.AI

Hive: A Multi-Agent Infrastructure for Algorithm- and Task-Level Scaling

Large language models are increasingly deployed as complex agentic systems that scale with task complexity. While prior work has extensively explored model- and system-level scaling, algorithm- and task-level scaling remain largely unaddressed, constraining the full potential of agentic systems. At the algorithm level, allocating additional inference-time computation can enhance workflow capacity but introduces cross-path redundancy: overlapping computations across multiple reasoning branches. At the task level, complex tasks can be decomposed into subproblems and delegated across multiple agents for improved scalability and parallelism. However, existing infrastructures' scheduling is unaware of the existence of multiple agents, missing opportunities to optimize resource allocation. We propose Hive, a multi-agent infrastructure that enables algorithm- and task-level scaling. Hive features a description frontend that captures per-agent behavior and supports test-time scaling algorithms. Leveraging this specification, our backend introduces two key mechanisms: Logits Cache that reuses intermediate logits across redundant sampling paths to mitigate cross-path redundancy at the algorithm level, and Agent-Aware Scheduling that efficiently allocates compute and KV-cache resources according to agent contributions at the task level. Experiments show that Logits Cache achieves an average speedup of

1.11\times

-

1.76\times

for re-sampling, and Agent-Aware Scheduling reduces the hotspot miss rate by

33\%

-

51\%

.

Zizhang Luo, Yuhao Luo, Youwei Xiao +3

Peking University Beijing, China

LLM Inference Efficiency Multi-Agent LLM Systems Multi-Agent System Optimization Inference-Time Scaling LLM Inference Acceleration Multi-Agent Systems

Apr 19, 2026q-fin.PM

Signal or Noise in Multi-Agent LLM-based Stock Recommendations?

We present the first portfolio-level validation of MarketSenseAI, a deployed multi-agent LLM equity system. All signals are generated live at each observation date, eliminating look-ahead bias. The system routes four specialist agents (News, Fundamentals, Dynamics, and Macro) through a synthesis agent that issues a monthly equity thesis and recommendation for each stock in its coverage universe, and we ask two questions: do its buy recommendations add value over both passive benchmarks and random selection, and what does the internal agent structure reveal about the source of the edge? On the S&P 500 cohort (19 months) the strong-buy equal-weight portfolio earns +2.18%/month against a passive equal-weight benchmark of +1.15% (approximating RSP), a +25.2% compound excess, and ranks at the 99.7th percentile of 10,000 Monte Carlo portfolios (p=0.003). The S&P 100 cohort (35 months) delivers a +30.5% compound excess over EQWL with consistent direction but formal significance not reached, limited by the small average selection of ~10 stocks per month. Non-negative least-squares projection of thesis embeddings onto agent embeddings reveals an adaptive-integration mechanism. Agent contributions rotate with market regime (Fundamentals leads on S&P 500, Macro on S&P 100, Dynamics acts as an episodic momentum signal) and this agent rotation moves in lockstep with both the sector composition of strong-buy selections and identifiable macro-calendar events, three independent views of the same underlying adaptation. The recommendation's cross-sectional Information Coefficient is statistically significant on S&P 500 (ICIR=+0.489, p=0.024). These results suggest that multi-agent LLM equity systems can identify sources of alpha beyond what classical factor models capture, and that the buy signal functions as an effective universe-filter that can sit upstream of any portfolio-construction process.

George Fatouros, Kostas Metaxas

Alpha Tensor Technologies

Multi-Agent LLM Systems Multi-Agent Collaboration Finance Algorithmic Trading

Apr 18, 2026cs.AI

GraphDC: A Divide-and-Conquer Multi-Agent System for Scalable Graph Algorithm Reasoning

Large Language Models (LLMs) have demonstrated strong potential for many mathematical problems. However, their performance on graph algorithmic tasks is still unsatisfying, since graphs are naturally more complex in topology and often require systematic multi-step reasoning, especially on larger graphs. Motivated by this gap, we propose GraphDC, a Divide-and-Conquer multi-agent framework for scalable graph algorithm reasoning. Specifically, inspired by Divide-and-Conquer design, GraphDC decomposes an input graph into smaller subgraphs, assigns each subgraph to a specialized agent for local reasoning, and uses a master agent to integrate the local outputs with inter-subgraph information to produce the final solution. This hierarchical design reduces the reasoning burden on individual agents, alleviates computational bottlenecks, and improves robustness on large graph instances. Extensive experiments show that GraphDC consistently outperforms existing methods on graph algorithm reasoning across diverse tasks and scales, especially on larger instances where direct end-to-end reasoning is less reliable.

Wenjin Li, Jiaming Cui

Department of Computer Science, Virginia Tech

Multi-Agent LLM Systems LLM Reasoning with Graphs Graph Reasoning Algorithmic Reasoning

Apr 18, 2026cs.CL

The Consensus Trap: Rescuing Multi-Agent LLMs from Adversarial Majorities via Token-Level Collaboration

Multi-agent large language model (LLM) architectures increasingly rely on response-level aggregation, such as Majority Voting (MAJ), to raise reasoning ceilings. However, in open environments, agents are highly susceptible to stealthy contextual corruption, such as targeted prompt injections. We reveal a critical structural vulnerability in current multi-agent systems: response-level aggregation collapses when corrupted agents form a local majority. Because voting aggregates fully-formed conclusions, it is blind to flawed intermediate logic. To overcome this systematic limitation, we propose the Token-Level Round-Robin (RR) Collaboration, where agents sequentially interleave generation within a shared auto-regressive context. We formalize this process as a discrete-time dynamical system, proving that token-level interleaving transitions aggregation from a brittle counting of final votes (a linear sum) to a dynamic, interwoven chain of logic (a non-linear operator product). Through this theoretical lens, we prove that the honest model's restorative pull can overpower adversarial corruptions, even when corrupted agents form a majority. We conduct an exhaustive empirical evaluation across diverse reasoning benchmarks and demonstrate that while MAJ collapses when corrupted agents reach a majority, RR maintains robust accuracy well beyond this critical threshold.

Jiayuan Liu, Shiyi Du, Weihua Du +2

Carnegie Mellon University · Foundations of Cooperative AI Lab (FOCAL) · Adelaide University

Multi-Agent LLM Systems Adversarial Robustness Multi-Agent Consensus Multi-Agent Reasoning

Apr 18, 2026cs.CL

StoryMI: Steerable Multi-Agent Therapeutic Dialogue Generation

Large language models (LLMs) can generate fluent dialogue, but prior works lack situational grounding, dynamic strategy control, and evaluation aligned with clinical standards in motivational interviewing (MI). We introduce StoryMI, a multi-LLM agent framework for controllable MI dialogue generation, where questionnaire-based client profiles are expanded into situational stories that provide narrative context for the dialogue. Therapist and client agents generate MI-coded utterances guided by MI codes selected by the interaction agent, while an interaction agent dynamically coordinates exchanges to control MI strategies during a multi-turn conversation. We propose a two-level evaluation protocol: lexical metrics and MI-specific measures of macro-level counseling strategies, alongside LLM-as-judge and human expert assessments. We construct a dataset of 6K simulated MI dialogues grounded in 1K questionnaire-story pairs, covering 12 MI codes and 13 symptom domains, and benchmark six open- and closed-source LLMs. Our results show that situational grounding and macro-level control can improve MI adherence and clinical plausibility, demonstrating the effectiveness of a structured multi-agent workflow for psychotherapy dialogue generation. We provide code and data for reproducibility.

Qingyu Meng, Min Chen, Dingming Liu +5

Vrije Universiteit Amsterdam · NII, Tokyo Institute of Technology

Mental Health Counseling Multi-Agent LLM Systems Emotional Support Conversation Mental Health Controllable Text Generation Conversational Agents

Apr 18, 2026cs.LG

Federation over Text: Insight Sharing for Multi-Agent Reasoning

Modern agents specialize in varying domains while there is no clear approach combining different domain skills. We propose a federated learning-like framework, Federation over Text (FoT), that enables multiple clients solving different tasks to collectively generate a shared library of metacognitive insights by iteratively federating their local reasoning processes without sharing actual problem instances. Instead of federation over gradients (e.g., as in distributed training), FoT operates at the semantic level without any gradient optimization or supervision signal. At each round, client LLM agents independently apply arbitrary local reasoning and self-improvement procedures to their own tasks and share the resulting reasoning traces with a central server. The server then aggregates, distills, and consolidates knowledge across tasks and domains into a shared insight library, which can be reused by current and future agents to improve their reasoning. Experiments show that FoT improves reasoning effectiveness and efficiency across real-world daily tasks, cross-domain collaboration, and research insight discovery, achieving an average performance gain of 11.9 absolute percentage points while reducing client-generated task-completion tokens by 5.5%.

Dixi Yao, Tahseen Rabbani, Manzil Zaheer +1

University of Chicago · Google DeepMind

Multi-Agent LLM Systems Multi-Agent Reasoning Federated Learning

Apr 17, 2026cs.AI

Debate as Reward: A Multi-Agent Reward System for Scientific Ideation via RL Post-Training

Large Language Models (LLMs) have demonstrated potential in automating scientific ideation, yet current approaches relying on iterative prompting or complex multi-agent architectures often suffer from hallucination or computational inefficiency. A critical bottleneck in applying Reinforcement Learning (RL) to this open-ended domain is reward hacking -- where models exploit imperfect evaluation proxies to maximize scores without producing genuine scientific innovation. To address these limitations, we propose an RL framework explicitly tailored for high-quality scientific idea generation. We propose the first multi-agent reward function designed to serve as a judge, decoupling methodological validation from implementation details while providing strict binary rewards that are robust to reward hacking. To effectively optimize against this sparse signal, we utilize an unbiased variant of Group Relative Policy Optimization to mitigate artificial length bias. We grounded our training in ICLR-320, a curated dataset of problem-solution pairs extracted from ICLR 2024 proceedings. Experiments demonstrate that our framework significantly outperforms state-of-the-art baselines across expert-evaluated metrics of novelty, feasibility, and effectiveness.

Moein Salimi, Babak Hosseini Mohtasham, Amin Aghakasiri +6

Sharif University of Technology · University of Tehran · Sharif University of Technology[cs.AI]

Reward Hacking RL for Language Models Multi-Agent LLM Systems Scientific Hypothesis Generation Large Language Model-Guided Scientific Discovery

Apr 17, 2026cs.MA

Agentic AI for Education: A Unified Multi-Agent Framework for Personalized Learning and Institutional Intelligence

Agentic Artificial Intelligence (AI) represents a paradigm shift from reactive systems to proactive, autonomous decision making frameworks. Existing AI-based educational systems remain fragmented and lack multi-level integration across stakeholders. This paper proposes the Agentic Unified Student Support System (AUSS), a novel multi-agent architecture integrating student-level personalization, educator-level automation, and institutional-level intelligence. The framework leverages Large Language Models (LLMs), reinforcement learning, predictive analytics, and rule-based reasoning. Experimental results demonstrate improvements in recommendation accuracy (92.4%), grading efficiency (94.1%), and dropout prediction (F1-score: 89.5%). The proposed system enables scalable, adaptive, and intelligent educational ecosystems.

Arya Mary K J, Deepthy K Bhaskar, Sinu T S +1

Department of Computer Engg, Govt. Model Engineering College,APJ Abdul Kalam Technological University, Kerala, India.

Multi-Agent LLM Systems Educational Technology AI in Education

Apr 17, 2026cs.SD

VoxMind: An End-to-End Agentic Spoken Dialogue System

Recent end-to-end spoken dialogue models enable natural interaction. However, as user demands become increasingly complex, models that rely solely on conversational abilities often struggle to cope. Incorporating agentic capabilities is therefore essential: by enabling tool use, these models can extend their knowledge boundaries and better solve real-world tasks. Yet, existing research has largely concentrated on core perception and generation, with comparatively limited exploration of such tool-augmented extensions. To bridge this gap, we present VoxMind, an integrated framework designed to equip end-to-end spoken dialogue models with comprehensive agentic abilities. Leveraging our curated 470-hour AgentChat dataset, we incorporate a "Think-before-Speak" mechanism, enabling the model to internalize structured reasoning as a critical prerequisite for planning and response generation. Furthermore, to mitigate latency bottlenecks caused by large-scale tool integration, we propose a Multi-Agent Dynamic Tool Management architecture. By asynchronously delegating retrieval tasks to an auxiliary agent aligned with the main model's reasoning trajectory, this system effectively decouples inference latency from toolset size. Experimental results confirm that VoxMind achieves significant improvements in agent performance: compared with strong baselines, the task completion rate increases from 34.88% to 74.57%, outperforming Gemini-2.5-Pro on spoken agent tasks while preserving general conversational quality. The source code and associated data are publicly available at https://github.com/MM-Speech/VoxMind.

Tianle Liang, Yifu Chen, Shengpeng Ji +7

Zhejiang University · China University of Petroleum-Beijing at Karamay · Xiamen University

Multi-Agent LLM Systems Spoken Dialogue Systems Tool-Using Agents Conversational Agents

Apr 17, 2026cs.MA

Conjunctive Prompt Attacks in Multi-Agent LLM Systems

Most LLM safety work studies single-agent models, but many real applications rely on multiple interacting agents. In these systems, prompt segmentation and inter-agent routing create attack surfaces that single-agent evaluations miss. We study \emph{conjunctive prompt attacks}, where a trigger key in the user query and a hidden adversarial template in one compromised remote agent each appear benign alone but activate harmful behavior when routing brings them together. We consider an attacker who changes neither model weights nor the client agent and instead controls only trigger placement and template insertion. Across star, chain, and DAG topologies, routing-aware optimization substantially increases attack success over non-optimized baselines while keeping false activations low. Existing defenses, including PromptGuard, Llama-Guard variants, and system-level controls such as tool restrictions, do not reliably stop the attack because no single component appears malicious in isolation. These results expose a structural vulnerability in agentic LLM pipelines and motivate defenses that reason over routing and cross-agent composition. Code is available at https://github.com/UCF-ML-Research/ConjunctiveAgents.

Nokimul Hasan Arif, Qian Lou, Mengxin Zheng

University of Central Florida, USA

Multi-Agent LLM Systems Adversarial Attacks on LLMs Prompt Injection Attacks Prompt Injection Attacks on AI Agents Multi-Agent System Security

Apr 16, 2026cs.AI

Autogenesis: A Self-Evolving Agent Protocol

Recent advances in LLM based agent systems have shown promise in tackling complex, long horizon tasks. However, existing agent protocols (e.g., A2A and MCP) under specify cross entity lifecycle and context management, version tracking, and evolution safe update interfaces, which encourages monolithic compositions and brittle glue code. We introduce Autogenesis Protocol (AGP), a self evolution protocol that decouples what evolves from how evolution occurs. Its Resource Substrate Protocol Layer (RSPL) models prompts, agents, tools, environments, and memory as protocol registered resources with explicit state, lifecycle, and versioned interfaces. Its Self Evolution Protocol Layer (SEPL) specifies a closed loop operator interface for proposing, assessing, and committing improvements with auditable lineage and rollback. Building on AGP, we present Autogenesis System (AGS), a self-evolving multi-agent system that dynamically instantiates, retrieves, and refines protocol-registered resources during execution. We evaluate AGS on multiple challenging benchmarks that require long horizon planning and tool use across heterogeneous resources. The results demonstrate consistent improvements over strong baselines, supporting the effectiveness of agent resource management and closed loop self evolution. The code is available at https://github.com/DVampire/Autogenesis.

Wentao Zhang, Zhe Zhao, Haibin Wen +4

Nanyang Technological University · Stanford University · City University of Hong Kong +1

Multi-Agent LLM Systems LLM Agent Self-Improvement LLM Agents Long-Horizon Agent Tasks Self-Evolving Agents

Apr 16, 2026cs.AI

El Agente Forjador: Task-Driven Agent Generation for Quantum Simulation

AI for science promises to accelerate the discovery process. The advent of large language models (LLMs) and agentic workflows enables the expediting of a growing range of scientific tasks. However, most of the current generation of agentic systems depend on static, hand-curated toolsets that hinder adaptation to new domains and evolving libraries. We present El Agente Forjador, a multi-agent framework in which universal coding agents autonomously forge, validate, and reuse computational tools through a four-stage workflow of tool analysis, tool generation, task execution, and iterative solution evaluation. Evaluated across 24 tasks spanning quantum chemistry and quantum dynamics on five coding agent setups, we compare three operating modes: zero-shot generation of tools per task, reuse of a curriculum-built toolset, and direct problem-solving with the coding agents as the baseline. We find that our tool generation and reuse framework consistently improves accuracy over the baseline. We also show that reusing a toolset built by a stronger coding agent can reduce API cost and substantially raises the solution quality for weaker coding agents. Case studies further demonstrate that tools forged for different domains can be combined to solve hybrid tasks. Taken together, these results show that LLM-based agents can use their scientific knowledge and coding capabilities to autonomously build reusable scientific tools, pointing toward a paradigm in which agent capabilities are defined by the tasks they are designed to solve rather than by explicitly engineered implementations.

Zijian Zhang, Aiwei Yin, Amaan Baweja +4

Department of Computer Science, University of Toronto, 40 St George St., Toronto, ON M5S 2E4, Canada · Vector Institute for Artificial Intelligence, W1140-108 College St., Schwartz Reisman Innovation2026 Campus, Toronto, ON M5G 0C6, Canada · NVIDIA, 431 King St W #6th, Toronto, ON M5V 1K4, Canada +5

Multi-Agent LLM Systems AI for Science Scientific Code Generation Tool-Using Agents LLM Agents

Apr 16, 2026cs.AI

Mind DeepResearch Technical Report

We present Mind DeepResearch (MindDR), an efficient multi-agent deep research framework that achieves leading performance with only ~30B-parameter models through a meticulously designed data synthesis and multi-stage training pipeline. The core innovation of MindDR lies in a collaborative three-agent architecture (Planning Agent, DeepSearch Agent, and Report Agent) and a four-stage agent-specialized training pipeline comprising SFT cold-start, Search-RL, Report-RL and preference alignment. With this regime, MindDR demonstrates competitive performance even with ~30B-scale models. Specifically, MindDR achieves 45.7% on BrowseComp-ZH, 42.8% on BrowseComp, 46.5% on WideSearch, 75.0% on xbench-DS, and 52.5 on DeepResearch Bench, outperforming comparable-scale open-source agent systems and rivaling larger-scale models. MindDR has been deployed as an online product in Li Auto. Furthermore, we introduce MindDR Bench, a curated benchmark of 500 real-world Chinese queries from our internal product user interactions, evaluated through a comprehensive multi-dimensional rubric system rather than relying on a single RACE metric. On MindDR Bench, MindDR achieves a state-of-the-art score of 51.8.

MindDR Team, Li Auto Inc

MindDR Team, Li Auto Inc

Multi-Agent LLM Systems Deep Research Agents AI Agent Benchmarks LLM Agent Training RL for LLM Agents

Apr 13, 2026cs.RO

M2HRI: An LLM-Driven Multimodal Multi-Agent Framework for Personalized Human-Robot Interaction

Multi-robot systems hold significant promise for social environments such as homes and hospitals, yet existing multi-robot systems often treat robots as functionally interchangeable, overlooking how distinct agent identities shape user perception and how such individuality changes the coordination requirements of multi-robot interaction. To address this, we introduce M2HRI, a multimodal multi-agent framework that models each robot as an identity-bearing agent through personality and long-term memory, together with a contextualized coordination mechanism that regulates agent participation. In a controlled user study (n = 105) in a multi-agent human-robot interaction (HRI) scenario, we found that most personality contrasts were distinguishable and consistently expressed. Long-term memory improved preference awareness and interaction naturalness, while contextualized coordination improved conversational flow, response appropriateness, and overlap avoidance. Together, these findings show that agent individuality and contextualized participation coordination play complementary roles in supporting coherent and socially appropriate multi-agent HRI. Project website available at https://project-m2hri.github.io/.

Shaid Hasan, Breenice Lee, Sujan Sarker +1

University of Virginia, VA, USA

Multi-Agent LLM Systems Multi-Agent Coordination Human-Robot Interaction Multimodal Agents Memory-Augmented Agents

Apr 13, 2026cs.SI

Collective Opinion Dynamics in Structured LLM Populations

Large Language Models are increasingly deployed as interacting agents in settings such as online platforms, recommendation systems, and multi-agent applications. Understanding the collective behaviors that emerge from their interactions is therefore increasingly crucial, especially as these behaviors may shape public opinion and contribute to polarization. In this work, we investigate how network structure and group composition shape the evolution of opinions in populations of LLM agents engaged in multi-round debates. We generate networks with controlled levels of homophily and varying group sizes, and perform ten independent runs per configuration. The results show that LLM agents exhibit patterns that are highly sensitive to network structure, relative group sizes, and to the model itself. We highlight that even within LLM populations, low homophily facilitates convergence by increasing opportunities for cross-group interaction, whereas higher homophily limits cross-group exposure and can preserve distinct opinion states, consistent with previous research. We further show that the choice of LLM substantially affects the resulting dynamics, with different models exhibiting different patterns of opinion updating under comparable network conditions. At the individual level, providing agents with information about their local neighborhood further modulates opinion transitions, revealing substantial differences between models in their sensitivity to local social context. Overall, our findings show that the collective dynamics of LLM populations arise from interactions among model-specific behavior, network structure, population composition, and local social information.

Erica Cau, Andrea Failla, Giulio Rossetti

Department of Computer Science, University of Pisa · ISTI-CNR

Multi-Agent LLM Systems Opinion Dynamics Multi-Agent Debate Agent-Based Modeling

Apr 12, 2026cs.LG

Diversity Without Fidelity: A Solver-Sampler Mismatch in Multi-Agent LLM Negotiation Simulation

Language models are increasingly used to simulate people: survey respondents, negotiators, stakeholders in policy exercises. In that role a model should reproduce how people plausibly behave, hesitating, conceding late, and settling for imperfect deals, rather than playing the best move. We call this the sampler role, in contrast to the solver role of finding the best move, and we test how the reasoning modes providers ship to strengthen models as solvers affect it. Our testbed is multi-party negotiation: five agents bargain over a regulation for fifteen turns, and unresolved issues are decided by an authority. Agents without a structured memory of the negotiation almost never reach agreement, whether reasoning is on or off: 314 of 315 such runs end with the authority deciding. What reasoning changes is how the failure looks. With reasoning enabled, one model family negotiates visibly, with varied moves, concessions in most runs, and a different path every time, yet still ends without agreement in fifteen runs of fifteen. Diversity checks would pass a model whose endings never change. Two further results show the task permits agreement: when agents write their own short running notes on the negotiation, agreement becomes the norm, while the same notes supplied ready-made change nothing; and hand-coded agents following textbook concession strategies agree in most runs under identical rules. Simulation pipelines should therefore vet models as samplers, on the distributions of outcomes they produce. Fidelity as a sampler must be tested on its own: solver strength is no guide to it, and switching on reasoning leaves it where it was.

Sandro Andric

Multi-Agent LLM Systems Multi-Agent Negotiation Automated Negotiation LLM Agent Evaluation Large Language Model-Based Social Simulation

Apr 6, 2026cs.AI

MMORF: A Multi-agent Framework for Designing Multi-objective Retrosynthesis Planning Systems

Multi-objective retrosynthesis planning is a critical chemistry task requiring dynamic balancing of quality, safety, and cost objectives. Language model-based multi-agent systems (MAS) offer a promising approach for this task: leveraging interactions of specialized agents to incorporate multiple objectives into retrosynthesis planning. We present MMORF, a framework for constructing MAS for multi-objective retrosynthesis planning. MMORF features modular agentic components, which can be flexibly combined and configured into different systems, enabling principled evaluation and comparison of different system designs. Using MMORF, we construct two representative MAS: MASIL and RFAS. On a newly curated benchmark consisting of 218 multi-objective retrosynthesis planning tasks, MASIL achieves strong safety and cost metrics on soft-constraint tasks, frequently Pareto-dominating baseline routes, while RFAS achieves a 48.6% success rate on hard-constraint tasks, outperforming state-of-the-art baselines. Together, these results show the effectiveness of MMORF as a foundational framework for exploring MAS for multi-objective retrosynthesis planning. Code and data are available at https://github.com/ninglab/MMORF.

Frazier N. Baker, Trieu Nguyen, Reza Averly +4

Department of Computer Science and Engineering, The Ohio State University · Department of Mathematics and Statistics, University of South Florida · Division of Medicinal Chemistry and Pharmacognosy, The Ohio State University +2

Multi-Agent LLM Systems Multi-Agent Planning Multi-Objective Optimization

Apr 3, 2026cs.LG

Marginal-Contribution Policy Gradients under Filtered Feedback for Multi-Agent LLMs

We develop a unified treatment of credit assignment for RL training in multi-agent LLM systems. We show that observed reward alone cannot distinguish an agent that determines it from one that never affects it, and that standard shared-reward training performs exact gradient ascent on each agent's private utility rather than system performance. Moreover, we prove no single scalar per agent can consistently account for joint performance once agents interact. We thus develop the unique background-dependent notion of marginal contribution satisfying natural consistency requirements. From it we derive gradient-correct marginal contribution training signals, identify them from filtered feedback, and optimally allocate a budget of exact counterfactual evaluations against learned-signal error. Instantiated in GRPO, our signal improves routed GSM8K accuracy over winner-take-all training at no extra generation cost.

Elai Ben-Gal, Stela Tong

Department of Mathematics Stanford University · Graduate School of Business Stanford University

Multi-Agent LLM Systems Multi-Agent System Optimization Credit Assignment Cooperative MARL Counterfactual Credit Assignment Fine-Grained Credit Assignment+1

Apr 1, 2026cs.AI

A Multi-Agent Human-LLM Collaborative Framework for Closed-Loop Scientific Literature Summarization

Scientific discovery is slowed by fragmented literature that requires excessive human effort to gather, analyze, and understand. AI tools, including autonomous summarization and question answering, have been developed to aid in understanding scientific literature. However, these tools lack the structured, multi-step approach necessary for extracting deep insights from scientific literature. Large Language Models (LLMs) offer new possibilities for literature analysis, but remain unreliable due to hallucinations and incomplete extraction. We introduce Elhuyar, a multi-agent, human-in-the-loop system that integrates LLMs, structured AI, and human scientists to extract, analyze, and iteratively refine insights from scientific literature. The framework distributes tasks among specialized agents for filtering papers, extracting data, fitting models, and summarizing findings, with human oversight ensuring reliability. The system generates structured reports with extracted data, visualizations, model equations, and text summaries, enabling deeper inquiry through iterative refinement. Deployed in materials science, it analyzed literature on tungsten under helium-ion irradiation, showing experimentally correlated exponential helium bubble growth with irradiation dose and temperature, offering insight for plasma-facing materials (PFMs) in fusion reactors. This demonstrates how AI-assisted literature review can uncover scientific patterns and accelerate discovery.

Maxwell J. Jacobson, Daniel Xie, Jackson Shen +6

Department of Computer Science, Purdue University, IN, USA · School of Materials Science and Engineering, Purdue University, IN, USA

Multi-Agent LLM Systems AI-Assisted Scientific Research Text Summarization Human-in-the-Loop AI LLM Information Extraction

Mar 30, 2026cs.CR

Kill-Chain Canaries: Stage-Level Tracking of Prompt Injection Across Attack Surfaces and Five Production LLMs

Multi-agent LLM systems now read documents, web pages and tool results on behalf of users, yet their resistance to prompt injection is usually reported as one number: did the attack succeed? We introduce a kill-chain canary method that plants a unique token in every injected payload and records the furthest of four stages it reaches (Exposed -> Persisted -> Relayed -> Executed), across 950 runs, five production LLMs, six attack surfaces, and five defense conditions. Exposure was 100% among runs that called the tool; the outcomes differ downstream. Claude Haiku 4.5 and Claude Sonnet 4.5 executed none of their 164 text-surface attacks, and in the text relay the canary token never appeared in a memory write (0/40); GPT-4o-mini executed 53% of its attacks. Four findings follow. (1) A Claude writer kept the canary token out of shared memory in every relay run we report; one cross-model pairing (Claude writer, GPT-4o-mini reader, n = 3) is consistent with this protecting the reader, and other pairings were not tested. (2) As readers, the Claude models executed 0/40 raw pre-seeded injections, but Claude Haiku 4.5 executed 2/3 injections relayed by GPT-4o-mini; whether relayed injections are harder to refuse than raw ones is an open question. (3) DeepSeek Chat went from 0/24 on pre-seeded memory to 8/8 on tool results, scenarios that also differ in task and payload format; white-text PDF payloads, invisible on the rendered page, succeeded at least as often as visible ones. (4) pi_detector and write_filter failed on channels they do not inspect, spotlighting failed on content it wraps, and write_filter blocked the PDF relay but not the text relay, a difference we cannot explain. Code and run logs are publicly released: https://github.com/KevinChunye/prompt_injection

Haochuan Kevin Wang, Zechen Zhang

Massachusetts Institute of Technology · University of Chicago

Multi-Agent LLM Systems Language Model Safety Evaluation Prompt Injection Defense Prompt Injection Attacks on LLMs Prompt Injection Attacks on AI Agents Prompt Injection Detection

Mar 27, 2026cs.CL

Ask or Assume? Uncertainty-Aware Clarification-Seeking in Coding Agents

As Large Language Model (LLM) agents are increasingly deployed in open-ended domains like software engineering, they frequently encounter underspecified instructions that lack crucial context. While human developers naturally resolve underspecification by asking clarifying questions, current agents are largely optimized for autonomous execution. In this work, we systematically evaluate the clarification-seeking abilities of LLM agents on an underspecified variant of SWE-bench Verified. We propose an uncertainty-aware multi-agent scaffold that decouples underspecification detection from code execution. Across both proprietary and open-weight frontier LLMs, our scaffold achieves a 69.40% task resolve rate, significantly outperforming a standard single-agent setup and closing the performance gap with agents operating on fully specified instructions. Furthermore, we find that the multi-agent system exhibits well-calibrated information-seeking behavior, conserving queries on simple tasks while proactively seeking information on more complex issues. These findings indicate that current models can be turned into proactive collaborators, where agents independently recognize when to ask questions to elicit missing information in real-world, underspecified tasks.

Nicholas Edwards, Sebastian Schuster

Faculty of Computer Science, University of Vienna, Vienna, Austria · UniVie Doctoral School Computer Science, University of Vienna, Vienna, Austria

Multi-Agent LLM Systems AI Coding Agents LLM Agents LLM Uncertainty Estimation

Mar 21, 2026cs.AI

Subjective-Graph LLM Agents for Simulating Uncertainty in Classroom Social Perception

Social actors do not observe a common social world: each individual forms judgments from a partial and potentially distorted view of the surrounding network. We study whether graph-local evidence and credibility-weighted communication can generate persistent distortions in perceived academic standing, even when agents repeatedly receive objective performance signals. We introduce a data-constrained multi-agent framework in which LLM agents operate through individualized subjective graphs that determine peer visibility, evidence access, and interaction opportunities. Agents exchange uncertainty-annotated assessments, evaluate message credibility, and maintain explicit Gaussian belief states updated through Bayesian fusion. We evaluate the framework on 12 middle-school classrooms comprising 482 students, using questionnaire-derived social information and six consecutive examinations. On the Social-Observed subset (n=419), collective ranking error increases from 0.066 \pm 0.008 to 0.124 \pm 0.009 across six epochs despite repeated exam-based anchoring. Ablations associate individualized visibility and LLM-based trust gating with more stable long-horizon behavior, while constrained retrieval primarily safeguards against global-information leakage. Compared with evaluated DeGroot configurations, the proposed framework achieves lower final ranking error; those DeGroot configurations exhibit near-zero terminal opinion diversity. These findings establish subjective-graph LLM agents as a mechanism-oriented framework for data-constrained simulated social perception. Code is available at https://anonymous.4open.science/r/Rashomonomon-0126.

Jinming Yang, Xinyu Jiang, Xinshan Jiao +1

Complex Lab, School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu, China · Department of Computer Science, City University of Hong Kong, Hong Kong · School of Health and Medical Technology, Chengdu Neusoft University, Chengdu,2026 China

Multi-Agent LLM Systems Opinion Dynamics AI in Education Computational Social Science Agent-Based Modeling

Mar 17, 2026cs.CL

Social Simulacra in the Wild: AI Agent Communities on Moltbook

As autonomous LLM-based agents increasingly populate social platforms, understanding the dynamics of AI-agent communities becomes essential for both communication research and platform governance. We present the first large-scale empirical comparison of AI-agent and human online communities, analyzing 73,899 Moltbook and 189,838 Reddit posts across five matched communities. Structurally, we find that Moltbook exhibits extreme participation inequality (Gini = 0.84 vs. 0.47) and high cross-community author overlap (33.8% vs. 0.5%). In terms of linguistic attributes, content generated by AI-agents is emotionally flattened, cognitively shifted toward assertion over exploration, and socially detached. These differences give rise to apparent community-level homogenization, but we show this is primarily a structural artifact of shared authorship. At the author level, individual agents are more identifiable than human users, driven by outlier stylistic profiles amplified by their extreme posting volume. As AI-mediated communication reshapes online discourse, our work offers an empirical foundation for understanding how multi-agent interaction gives rise to collective communication dynamics distinct from those of human communities.

Agam Goyal, Olivia Pal, Hari Sundaram +2

Siebel School of Computing and Data Science University of Illinois Urbana-Champaign, Urbana, IL, USA

Multi-Agent LLM Systems Multi-Agent Communication Agentic AI

Mar 9, 2026cs.CL

EvoScientist: Towards Multi-Agent Evolving AI Scientists for End-to-End Scientific Discovery

The increasing adoption of Large Language Models (LLMs) has enabled AI scientists to perform complex end-to-end scientific discovery tasks requiring coordination of specialized roles, including idea generation and experimental execution. However, most state-of-the-art AI scientist systems rely on static, hand-designed pipelines and fail to adapt based on accumulated interaction histories. As a result, these systems overlook promising research directions, repeat failed experiments, and pursue infeasible ideas. To address this, we introduce EvoScientist, an evolving multi-agent AI scientist framework that continuously improves research strategies through persistent memory and self-evolution. EvoScientist comprises three specialized agents: a Researcher Agent (RA) for scientific idea generation, an Engineer Agent (EA) for experiment implementation and execution, and an Evolution Manager Agent (EMA) that distills insights from prior interactions into reusable knowledge. EvoScientist contains two persistent memory modules: (i) an ideation memory, which summarizes feasible research directions from top-ranked ideas while recording previously unsuccessful directions; and (ii) an experimentation memory, which captures effective data processing and model training strategies derived from code search trajectories and best-performing implementations. These modules enable the RA and EA to retrieve relevant prior strategies, improving idea quality and code execution success rates over time. Experiments show that EvoScientist outperforms 7 open-source and commercial state-of-the-art systems in scientific idea generation, achieving higher novelty, feasibility, relevance, and clarity via automatic and human evaluation. EvoScientist also substantially improves code execution success rates through multi-agent evolution, demonstrating persistent memory's effectiveness for end-to-end scientific discovery.

Yougang Lyu, Xi Zhang, Yuyue Zhao +11

Huawei Technologies Co., Ltd. · Vrije Universiteit Amsterdam

Multi-Agent LLM Systems AI Agents for Scientific Discovery Scientific Code Generation Memory-Augmented Agents Self-Evolving Agents

Mar 6, 2026cs.LG

The Trace Is the State: Exact Credit Assignment for LLM Agent Teams

Credit assignment for a team of LLM agents, what each message was worth, has mostly been treated as prediction: a learned critic, a trajectory-level score, or an agent ablation stands in for the counterfactual that defines credit, which is rarely run. Teams that communicate through a shared context are different: when everything a downstream agent reads is written into the trace, the trace is the state, and that counterfactual can be executed. A credit signal can then be judged as any estimator is: by bias, variance, and agreement with an independent reference. C3, credit assignment by counterfactual continuation, substitutes one message at a decision point and continues the run to the terminal reward, so its credit is unbiased, exact up to Monte Carlo error. Given the sampled alternatives, that error's variance follows a derived law with no term for the number of agents, and the observed noise follows the law on 6 workflows of 2 to 10 decision points. On the 2-agent chain, from 4 continuations, C3 ranks alternatives at 0.69 rank correlation against a 16-continuation reference, near the 0.73 at which 2 such references agree; a critic trained on the same rollouts reaches 0.29. Used as the advantage in training a 2-agent team, C3 beats MAGRPO, the stronger baseline in our comparison, and spends 37% fewer training tokens than MAPPO, since only the messages downstream of a decision point are regenerated. When the trace is the state, credit need not be predicted; it can be exact.

Yanjun Chen, Yirong Sun, Hanlin Wang +5

Eastern Institute of Technology · The Hong Kong Polytechnic University · Harbin Institute of Technology

Credit Assignment Counterfactual Credit Assignment LLM Agent Training Multi-Agent LLM Systems Multi-Agent Reinforcement Learning

Mar 2, 2026cs.AI

Exploring Plan Space through Conversation: An Agentic Framework for LLM-Mediated Explanations in Planning

When automating plan generation for a real-world sequential decision problem, the goal is often not to replace the human planner, but to facilitate an iterative reasoning and elicitation process, where the human's role is to guide the AI planner according to their preferences and expertise. In this context, explanations that respond to users' questions are crucial to improve their understanding of potential solutions and increase their trust in the system. To enable natural interaction with such a system, we present a multi-agent Large Language Model (LLM) architecture that is agnostic to the explanation framework and enables user- and context-dependent interactive explanations. We also describe an instantiation of this framework for goal-conflict explanations, which we use to conduct a user study comparing the LLM-powered interaction with a baseline template-based explanation interface.

Guilhem Fouilhé, Rebecca Eifler, Antonin Poché +2

IRIT, Toulouse, France · LAAS-CNRS, Toulouse, France · IRT Saint Exupery, Toulouse, France +2

Multi-Agent LLM Systems Explainable Artificial Intelligence Human-in-the-Loop AI

Mar 1, 2026cs.AI

HVR-Met: A Hypothesis-Verification-Replanning Agentic System for Extreme Weather Diagnosis

While deep learning-based weather forecasting paradigms have made significant strides, addressing extreme weather diagnostics remains a formidable challenge. This gap exists primarily because the diagnostic process demands sophisticated multi-step logical reasoning, dynamic tool invocation, and expert-level prior judgment. Although agents possess inherent advantages in task decomposition and autonomous execution, current architectures are still hampered by critical bottlenecks: inadequate expert knowledge integration, a lack of professional-grade iterative reasoning loops, and the absence of fine-grained validation and evaluation systems for complex workflows under extreme conditions. To this end, we propose HVR-Met, a multi-agent meteorological diagnostic system characterized by the deep integration of expert knowledge. Its central innovation is the ``Hypothesis-Verification-Replanning'' closed-loop mechanism, which facilitates sophisticated iterative reasoning for anomalous meteorological signals during extreme weather events. To bridge gaps within existing evaluation frameworks, we further introduce a novel benchmark focused on atomic-level subtasks. Experimental evidence demonstrates that the system excels in complex diagnostic scenarios.

Shuo Tang, Jiadong Zhang, Gengxian Zhou +11

MAIS, Institute of Automation, Chinese Academy of Sciences · School of Artificial Intelligence, University of Chinese Academy of Sciences · Zhongguancun Academy, Beijing +3

Multi-Agent LLM Systems AI Agent Benchmarks

Feb 25, 2026cs.AI

Power and Limitations of Aggregation in Compound AI Systems

When designing compound AI systems, a common approach is to query multiple copies of the same model and aggregate the responses to produce a synthesized output. Given the homogeneity of these models, this raises the question of whether aggregation unlocks access to a greater set of outputs than querying a single model. In this work, we investigate the power and limitations of aggregation within a stylized principal-agent framework. This framework models how the system designer can partially steer each agent's output through its reward function specification, but still faces limitations due to prompt engineering ability and model capabilities. Our analysis uncovers three natural mechanisms -- feasibility expansion, support expansion, and binding set contraction -- through which aggregation expands the set of outputs that are elicitable by the system designer. We prove that any aggregation operation must implement one of these mechanisms in order to be elicitability-expanding, and that strengthened versions of these mechanisms provide necessary and sufficient conditions that fully characterize elicitability-expansion. Finally, we provide an empirical illustration of our findings for LLMs deployed in a toy reference-generation task. Altogether, our results take a step towards characterizing when compound AI systems can overcome limitations in model capabilities and in prompt engineering.

Nivasini Ananthakrishnan, Meena Jagadeesan

UC Berkeley · Stanford University

Multi-Agent LLM Systems

Feb 23, 2026cs.GT

A General Equilibrium Theory of Orchestrated AI Agent Systems

We establish a general equilibrium theory for systems of large language model (LLM) agents operating under centralized orchestration. The framework is a production economy in the sense of Arrow-Debreu (1954), extended to infinite-dimensional commodity spaces following Bewley (1972). Each LLM agent is modeled as a firm whose production set Y a

\subset

H = L 2 ([0, T ], R R ) represents the feasible metric trajectories determined by its frozen model weights. The orchestrator is the consumer, choosing a routing policy over the agent DAG to maximize system welfare subject to a budget constraint evaluated at functional prices p

\in

H A . These prices-elements of the Hilbert dual of the commodity space-assign a shadow value to each metric of each agent at each instant. We prove, via Brouwer's theorem applied to a finitedimensional approximation V K

\subset

H, that every such economy admits at least one general equilibrium (p * , y * ,

π

* ). A functional Walras' law holds as a theorem: the value of functional excess demand is zero for all prices, as a consequence of the consumer's budget constraint-not by construction. We further establish Pareto optimality (First Welfare Theorem), decentralizability of Pareto optima (Second Welfare Theorem), and uniqueness with geometric convergence under a contraction condition (Banach). The orchestration dynamics constitute a Walrasian t{â}tonnement that converges globally under the contraction condition, unlike classical t{â}tonnement (Scarf, 1960). The framework admits a DSGE interpretation with SLO parameters as policy rates.

Jean-Philippe Garnier

Br.AI.K · BrainiaK

Multi-Agent LLM Systems Multi-Agent Coordination LLM Agent Orchestration

Feb 18, 2026cs.MA

Evaluating Collective Behaviour of Hundreds of LLM Agents

LLM-powered AI assistants acting on behalf of users can produce poor collective outcomes at scale. We introduce a framework for evaluating their emergent behaviour in social dilemmas, applied to three iterated games (Public Goods, Collective Risk, Common Pool Resource). We prompt each model to produce a natural-language strategy, then have the same model translate it into code. This aims to isolate strategic reasoning from input-parsing, enables pre-deployment inspection, and scales to populations of hundreds of agents. We propose three analyses: behavioural fingerprinting via exhaustive evaluation over opponent histories; self-play robustness across mixtures of a model's strategies with either a Selfish or Collective disposition; and cultural evolution under payoff-biased imitation. Applied to three state-of-the-art LLMs, we find substantial cross-model differences in self-play welfare, and that cultural evolution converges to low-welfare, Selfish-dominant equilibria in larger groups.

Richard Willis, Jianing Zhao, Yali Du +1

King’s College London London, United Kingdom · King’s College London · Turing Institute +1

Social Dilemmas Multi-Agent LLM Systems Cultural Evolution Game Theory LLM Agent Evaluation Large Language Model-Based Social Simulation+1