Watch Every Step! LLM Agent Learning via Iterative Step-Level Process Refinement.
Weimin XiongYifan SongXiutian ZhaoWenhao WuXun WangKe WangCheng LiWei PengSujian LiPublished in: CoRR (2024)
Keyphrases
- learning process
- learning systems
- learning phase
- action selection
- knowledge level
- feedback loop
- knowledge acquisition
- autonomous agents
- learning algorithm
- iterative process
- goal directed
- learning stage
- unsupervised learning
- multi agent systems
- decision making
- iterative learning
- refinement process
- refinement step
- neural network
- learning agent
- human agent
- reasoning process
- learning mechanism
- agent model
- learning problems
- dynamic environments
- cooperative
- multi agent
- reinforcement learning