Two Heads are Better Than One: Test-time Scaling of Multi-agent Collaborative Reasoning

2025-04-14 · Can Jin, Hongwu Peng, Qixin Zhang, Yujin Tang, Dimitris N. Metaxas, Tong Che

Multi-agent systems (MAS) built on large language models (LLMs) offer a promising path toward solving complex, real-world tasks that single-agent systems often struggle to manage. While recent advancements in test-time scaling (TTS) have significantly improved single-agent performance on challenging reasoning tasks, how to effectively scale collaboration and reasoning in MAS remains an open question. In this work, we introduce an adaptive multi-agent framework designed to enhance collaborative reasoning through both model-level training and system-level coordination. We construct M500, a high-quality dataset containing 500 multi-agent collaborative reasoning traces, and fine-tune Qwen2.5-32B-Instruct on this dataset to produce M1-32B, a model optimized for multi-agent collaboration. To further enable adaptive reasoning, we propose a novel CEO agent that dynamically manages the discussion process, guiding agent collaboration and adjusting reasoning depth for more effective problem-solving. Evaluated in an open-source MAS across a range of tasks-including general understanding, mathematical reasoning, and coding-our system significantly outperforms strong baselines. For instance, M1-32B achieves 12% improvement on GPQA-Diamond, 41% on AIME2024, and 10% on MBPP-Sanitized, matching the performance of state-of-the-art models like DeepSeek-R1 on some tasks. These results highlight the importance of both learned collaboration and adaptive coordination in scaling multi-agent reasoning. Code is available at https://github.com/jincan333/MAS-TTS

📄 PDF Abstract BibTeX arXiv:2504.09772

Code (1)

jincan333/MAS-TTS 공식 구현 pytorch

Tasks

Mathematical Reasoningmbpp

Methods 이 논문이 사용한 방법론

MAS This optimizer mix ADAM and SGD creating the MAS optimizer.

Distilled Pretraining: A modern lens of Data, In-Context Learning and Test-Time Scaling

2025-09-01 · Sachin Goyal, David Lopez-Paz, Kartik Ahuja arxiv

In the past year, distillation has seen a renewed prominence in large language model (LLM) pretraining, exemplified by the Llama-3.2 and Gemma model families. While distillation has historically been shown to improve sta…

Large-Scale High-Quality 3D Gaussian Head Reconstruction from Multi-View Captures

2026-05-05 · Evangelos Ntavelis, Sean Wu, Mohamad Shahbazi, Fabio Maninchedda 외 arxiv

We propose HeadsUp, a scalable feed-forward method for reconstructing high-quality 3D Gaussian heads from large-scale multi-camera setups. Our method employs an efficient encoder-decoder architecture that compresses inpu…

Mutually-Constrained Monotonic Multihead Attention for Online ASR

2021-03-26 · Jaeyun Song, Hajin Shim, Eunho Yang

Despite the feature of real-time decoding, Monotonic Multihead Attention (MMA) shows comparable performance to the state-of-the-art offline methods in machine translation and automatic speech recognition (ASR) tasks. How…

Automatic Speech Recognition Automatic Speech Recognition (ASR)Machine Translation speech-recognition+2

Toward Localizing and Repairing Bias in Transformer Attention Heads

2026-07-14 · Sigma Jahan arxiv

Transformer language models are increasingly used as software components, yet biased outputs remain difficult to localize and repair inside the model. Existing fairness testing and repair methods largely operate at the i…

Scaling Reasoning Hop Exposes Weaknesses: Demystifying and Improving Hop Generalization in Large Language Models

2026-01-29 · Zhaoyi Li, Jiatong Li, Gangwei Jiang, Linqi Song 외 arxiv

Chain-of-thought (CoT) reasoning has become the standard paradigm for enabling Large Language Models (LLMs) to solve complex problems. However, recent studies reveal a sharp performance drop in reasoning hop generalizati…