arXiv AI September 4, 2026 By Xuanfa Jin, Zhijian Ma, Yongcheng Zeng, Xinyu Cui, Haifeng Zhang, Jun Wang

Remember and Reweight: Enhancing Multi-Agent Debate with Experience Memory and Confidence Estimation

The paper introduces R$^2$-MAD, a framework that enhances multi-agent debate by giving agents an experience memory from past debates. It uses a debate-state-aware retrieval policy to adjust concept priors based on current consensus, and derives confidence weights from retrieved experiences to modulate peer influence. Experiments demonstrate consistent improvements over existing single-agent and MAD baselines.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

arXiv AI

Jun 2

Demystifying Multi-Agent Debate: The Role of Confidence and Diversity

arXiv:2601. 19921v2 Announce Type: replace-cross Abstract: Multi-agent debate (MAD) is widely used to improve large language model (LLM) performance through test-time scaling, yet recent work shows that vanilla MAD often underperforms simple majority vote despite higher computational cost.

By Xiaochen Zhu, Caiqi Zhang, Yizhou Chi, Tom Stafford, Nigel Collier, Andreas Vlachos

llms agents benchmarks

arXiv Machine Learning

Jul 15

When and Why Does Multi-Agent Debate Fail and Does It Really Underperform?

arXiv:2510. 20963v2 Announce Type: replace Abstract: Multi-agent debate (MAD) was proposed as a promising approach for ensembling the wisdom of multiple large language models (LLMs) to improve reasoning and provide effective supervision to superhuman LLMs.

By Yongqiang Chen, Gang Niu, James Cheng, Bo Han, Masashi Sugiyama

llms agents benchmarks

arXiv AI

Jul 16

Too Polite to Disagree: Understanding Sycophancy Propagation in Multi-Agent Systems

arXiv:2604. 02668v2 Announce Type: replace-cross Abstract: Large language models (LLMs) often exhibit sycophancy: agreement with user stance even when it conflicts with the model's opinion.

By Vira Kasprova, Amruta Parulekar, Abdulrahman AlRabah, Krishna Agaram, Ritwik Garg, Sagar Jha, Nimet Beyza Bozdag, Dilek Hakkani-Tur

llms agents

arXiv AI

Jun 12

Multiagent Protocols with Aggregated Confidence Signals

arXiv:2606. 13591v1 Announce Type: new Abstract: Confidence is used for reliability, oversight, and a range of downstream decision tasks in Natural Language Processing (NLP), yet no existing method produces or evaluates a confidence for the output of a multiagent system.

By Ali Elahi, Barbara Di Eugenio

agents nlp benchmarks

arXiv AI

Jul 31

Multi-Agent Debate Strategies: Survey, Taxonomy, and Challenges

arXiv:2607. 26212v1 Announce Type: cross Abstract: Multi-Agent Debate (MAD) is a promising paradigm for improving the accuracy and robustness of Large Language Model (LLM)-based agentic systems.

By Quim Motger, Marc Oriol, Jordi Marco, Xavier Franch

llms agents benchmarks

arXiv AI

Aug 5

Where Reasoning Diverges: Localized Multi-Agent Debate for Multi-Hop Question Answering

arXiv:2608. 01463v2 Announce Type: replace Abstract: Multi-agent debate commonly exchanges complete rationales even when disagreements concern only a few intermediate claims.

By Weijun Gao, Xiang Ding, Haoyang Liu, Tiancheng Xing

agents nlp benchmarks

One email a morning, machine-written

Related stories

Demystifying Multi-Agent Debate: The Role of Confidence and Diversity

When and Why Does Multi-Agent Debate Fail and Does It Really Underperform?

Too Polite to Disagree: Understanding Sycophancy Propagation in Multi-Agent Systems

Multiagent Protocols with Aggregated Confidence Signals

Multi-Agent Debate Strategies: Survey, Taxonomy, and Challenges

Where Reasoning Diverges: Localized Multi-Agent Debate for Multi-Hop Question Answering