Enhancing LLM Reasoning with Multi-Path Collaborative Reactive and Reflection agents

2024-12-31

AI Summary

Large language models (LLMs) suffer from low accuracy and reasoning degeneration in complex scientific reasoning tasks. To address this, we propose RR-MP—a zero-shot or few-shot multi-path parallel reasoning framework wherein each path comprises a reactive agent and a reflective agent operating synergistically; this is the first multi-path reasoning architecture driven by a dual-agent mechanism. RR-MP requires no model fine-tuning and improves reasoning quality solely through iterative inter-agent dialogues and cross-path summary fusion. Evaluated on moral judgment, university-level physics, and mathematical reasoning benchmarks, RR-MP substantially outperforms state-of-the-art baselines, demonstrating superior accuracy, robustness, and generalization in high-difficulty scientific reasoning. Its core innovations are: (1) a reaction–reflection dual-agent collaboration paradigm that enables dynamic, self-correcting inference; and (2) a training-free multi-path ensemble mechanism that aggregates diverse reasoning trajectories without parameter updates.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Multiagent Systems: Multiagent Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Agents have demonstrated their potential in scientific reasoning tasks through large language models. However, they often face challenges such as insufficient accuracy and degeneration of thought when handling complex reasoning tasks, which impede their performance. To overcome these issues, we propose the Reactive and Reflection agents with Multi-Path Reasoning (RR-MP) Framework, aimed at enhancing the reasoning capabilities of LLMs. Our approach improves scientific reasoning accuracy by employing a multi-path reasoning mechanism where each path consists of a reactive agent and a reflection agent that collaborate to prevent degeneration of thought inherent in single-agent reliance. Additionally, the RR-MP framework does not require additional training; it utilizes multiple dialogue instances for each reasoning path and a separate summarizer to consolidate insights from all paths. This design integrates diverse perspectives and strengthens reasoning across each path. We conducted zero-shot and few-shot evaluations on tasks involving moral scenarios, college-level physics, and mathematics. Experimental results demonstrate that our method outperforms baseline approaches, highlighting the effectiveness and advantages of the RR-MP framework in managing complex scientific reasoning tasks.

Problem

• Large Language Models
• Complex Scientific Reasoning
• Accuracy and Cognitive Quality

Innovation

• Multi-Agent Reasoning
• Reactive and Reflective Agents
• Enhanced Scientific Inference