AI Dynamics

Global AI News Aggregator

Language Agent Tree Search: Balancing Exploration and Reward

3/ Language Agent Tree Search (LATS) Algorithm combines reflection and Monte-Carlo tree search to find the “best” trajectory for a task. The search process balances exploration and expected reward. It has 4 main steps: 1. Select: pick the best next actions based on the

→ View original post on X — @langchain,

Commentaires

Leave a Reply

Your email address will not be published. Required fields are marked *