mlsanityData to 14 Oct 2026

Learning Adaptive Deliberation Control for Reliable Recognition with Multi-Agent Language Models

arXiv 2610.17840 ↗First posted 13 Oct 2026cs.AIcited by 0

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks also fits: Reinforcement learning for LLM agents

Abstract (arXiv)

Multi-agent deliberation can improve recognition by criticizing and revising candidate answers, but fixed controls cannot respond to changes in disagreement or reasoning quality. We present an adaptive multi-agent controller (ADMA) that learns continuous policies for proposal, challenge, arbitration, and coordination while keeping the task model frozen. Shared trace features describe quality, agreement, and progress, while role-specific features support decentralized control during inference. One controller trained on legal and financial reasoning is transferred without policy updates to four textual and multimodal benchmarks and two successive task-model versions. With Qwen3.5 held fixed, ADMA-Full improves accuracy by 6.2-14.2 percentage points over Direct and by 2.8-3.0 points over ADMA-Fixed. Paired bootstrap ranges for the Full-Fixed comparison are positive on three tasks. Outcome decompositions show that recoveries exceed introduced errors, while process assessment records fewer false-consensus events and more supported correct outcomes. Transfer to Qwen3.6 remains task dependent, which bounds the generality of the learned policy. Median token and latency ratios are 14.3-18.7 and 7.8-16.5 relative to Direct. These results show that adaptive deliberation can improve fixed-model recognition when its task-dependent gains justify the additional computation.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 27 of its 40 references that are in the map. First posted on arXiv 13 Oct 2026.

Builds on

27 of 40 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025AI Agents vs. Agentic AI: A Conceptual Taxonomy, Applications and Challenges15 May
  2. 2024Improving Multi-Agent Debate with Sparse Communication Topology17 Jun
  3. 2023A Survey on Large Language Model based Autonomous Agents22 Aug

Other references · 24

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  3. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  4. 2023G-Eval: NLG Evaluation using GPT-4 with Better Human Alignmentcited by 2,276
  5. 2021Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processingcited by 2,234
  6. 2023MMBench: Is Your Multi-modal Model an All-around Player?cited by 2,141
  7. 2023Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debatecited by 1,011
  8. 2024Are We on the Right Way for Evaluating Large Vision-Language Models?cited by 914
  9. 2023SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Modelscited by 810
  10. 2021FinQA: A Dataset of Numerical Reasoning over Financial Datacited by 562
  11. 2021Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Surveycited by 463
  12. 2023MedAgents: Large Language Models as Collaborators for Zero-shot Medical Reasoningcited by 316
12 more
  1. 2021LexGLUE: A Benchmark Dataset for Legal Language Understanding in Englishcited by 310
  2. 2023ReConcile: Round-Table Conference Improves Reasoning via Consensus among Diverse LLMscited by 309
  3. 2021When Does Pretraining Help? Assessing Self-Supervised Learning for Law and the CaseHOLD Datasetcited by 201
  4. 2024Multi-Step Reasoning with Large Language Models, a Surveycited by 143
  5. 2024A Survey on Uncertainty Quantification of Large Language Models: Taxonomy, Open Research Challenges, and Future Directionscited by 121
  6. 2023Can ChatGPT Defend its Belief in Truth? Evaluating LLM Reasoning via Debatecited by 94
  7. 2024Out-of-Distribution Detection: A Task-Oriented Survey of Recent Advancescited by 26
  8. 2021Taxonomy of Machine Learning Safety: A Survey and Primercited by 25
  9. 2020Explainable Goal-Driven Agents and Robots -- A Comprehensive Reviewcited by 24
  10. 2024Enhancing textual textbook question answering with large language models and retrieval augmented generationcited by 14
  11. 2024Explaining Predictive Uncertainty by Exposing Second-Order Effectscited by 10
  12. 2023Survey on Leveraging Uncertainty Estimation Towards Trustworthy Deep Neural Networks: The Case of Reject Option and Post-training Processingcited by 7

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026ECMR: Evidence-Aware Deliberation Control for Multi-Agent LLM Reasoning
  2. 2025Multi-Agent Collaborative Intelligence: Dual-Dial Control for Reliable LLM Reasoning
  3. 2025Multi-Agent Debate for LLM Judges with Adaptive Stability Detection
  4. 2025Self-Improvement of Language Models by Post-Training on Multi-Agent Debate
  5. 2026Auditing Multi-Agent LLM Reasoning Trees Outperforms Majority Vote and LLM-as-Judge
  6. 2026JudgePanel: A Compact Judge with Panel Deliberation via Adaptive Multi-Reward Reinforcement Learning
  7. 2026Meta-Moderator: Empowering Multi-Agent Debate with Meta-Cognition
  8. 2025From Efficiency to Adaptivity: A Deeper Look at Adaptive Reasoning in Large Language Models