Don't throw away your value model! Generating more preferable text with Value-Guided Monte-Carlo Tree Search decoding

arXiv:2309.15028v3 Announce Type: replace-cross
Abstract: Inference-time search algorithms such as Monte-Carlo Tree Search (MCTS) may seem unnecessary when generating natural language text based on state-of-the-art reinforcement learning such as Proximal Policy Optimization (PPO). In this paper, we demonstrate that it is possible to get extra mileage out of PPO by integrating MCTS on top. The key idea is not to throw out the value network, a byproduct of PPO training for evaluating partial output sequences, when decoding text out of …

abstract algorithms art arxiv cs.ai cs.cl cs.lg decoding inference language monte-carlo natural natural language optimization paper policy ppo reinforcement reinforcement learning search search algorithms state text tree type value