Reducing Cognitive Overhead in Tool Use via Multi-Small-Agent Reinforcement Learning

arXiv:2508.08882 · cs.AI · Submitted 2025-08-12 · Read on arXiv

cs.AI

Submitted: 2025-08-12

Updated: 2026-10-01

Code: https://github.com/dayuwang401/MSARL-

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Reinforcement Learning — 3× in this paper · explained in 4 episodes
• Accuracy — 1× in this paper · explained in 1 episode
• Decoupling — 1× in this paper · explained in 2 episodes
• Multi-Agent System — 1× in this paper · explained in 7 episodes

Sources

• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• ToRA: A Tool-Integrated Reasoning Agent for Mathematical Problem Solving
• LLM Multi-Agent Systems: Challenges and Open Problems
• OpenAI o1 System Card
• Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
• MACM: Utilizing a Multi-Agent System for Condition Mining in Solving Complex Mathematical Problems
• ToRL: Scaling Tool-Integrated RL
• Let's Verify Step by Step
• Qwen3 Technical Report
• Proximal Policy Optimization Algorithms
• Multi-Agent Collaboration Mechanisms: A Survey of LLMs
• Mixture-of-Agents Enhances Large Language Model Capabilities
• MathCoder: Seamless Code Integration in LLMs for Enhanced Mathematical Reasoning
• OpenAgents: An Open Platform for Language Agents in the Wild
• Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement
• Auto-GPT for Online Decision Making: Benchmarks and Additional Opinions
• StepTool: Enhancing Multi-Step Tool Usage in LLMs via Step-Grained Reinforcement Learning
• SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild
• G-Designer: Architecting Multi-agent Communication Topologies via Graph Neural Networks
• Critique-GRPO: Advancing LLM Reasoning with Natural Language and Numerical Feedback

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems