Get Started
Home
Topics
Search
Library
Research questionHow can LLM agents jointly adapt reasoning policies and hierarchical skill libraries during reinforcement learning?Skill libraries can improve reinforcement-learning efficiency, but fixed meta-skill workflows and separate skill evolution can prevent skills from adapting to the agent's changing reasoning policy. The central difficulty is coordinating updates to reusable skills with the behavior that selects and applies them.
AI
AI Agents
Machine Learning
Multi-agent Systems
Reasoning
Reinforcement Learning
Latest papersRecent research connected to this question, newest first.CoSkill: Joint Reinforcement Learning of Reasoning and Meta-Skill Agents for Hierarchical Skill EvolutionThe source studies a cooperative Reasoning Agent and Meta-Skill Agent sharing one backbone and jointly learning over a hierarchical skill library. The Reasoning Agent uses retrieved task and step skills, while task performance informs skill refinement; evidence comes from experiments on ALFWorld and WebShop.research paper · Sep 4, 2026
Related questions
How can LLM agents reuse procedural skills across diverse long-horizon tasks without generic or instance-bound memory?How can LLM agents generalize to unseen tasks without directly fine-tuning their policies?How can LLM agents retrieve relevant skills from large, noisy libraries under context and latency constraints?How can LLM agents reuse execution traces without losing temporal and outcome-dependent behavior?