Targeted Search Control in AlphaZero for Effective Policy Improvement

02/23/2023
by   Alexandre Trudeau, et al.
0

AlphaZero is a self-play reinforcement learning algorithm that achieves superhuman play in chess, shogi, and Go via policy iteration. To be an effective policy improvement operator, AlphaZero's search requires accurate value estimates for the states appearing in its search tree. AlphaZero trains upon self-play matches beginning from the initial state of a game and only samples actions over the first few moves, limiting its exploration of states deeper in the game tree. We introduce Go-Exploit, a novel search control strategy for AlphaZero. Go-Exploit samples the start state of its self-play trajectories from an archive of states of interest. Beginning self-play trajectories from varied starting states enables Go-Exploit to more effectively explore the game tree and to learn a value function that generalizes better. Producing shorter self-play trajectories allows Go-Exploit to train upon more independent value targets, improving value training. Finally, the exploration inherent in Go-Exploit reduces its need for exploratory actions, enabling it to train under more exploitative policies. In the games of Connect Four and 9x9 Go, we show that Go-Exploit learns with a greater sample efficiency than standard AlphaZero, resulting in stronger performance against reference opponents and in head-to-head play. We also compare Go-Exploit to KataGo, a more sample efficient reimplementation of AlphaZero, and demonstrate that Go-Exploit has a more effective search control strategy. Furthermore, Go-Exploit's sample efficiency improves when KataGo's other innovations are incorporated.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
05/30/2020

Manipulating the Distributions of Experience used for Self-Play Learning in Expert Iteration

Expert Iteration (ExIt) is an effective framework for learning game-play...
research
08/03/2022

Supervised and Reinforcement Learning from Observations in Reconnaissance Blind Chess

In this work, we adapt a training approach inspired by the original Alph...
research
02/10/2020

Provable Self-Play Algorithms for Competitive Reinforcement Learning

Self-play, where the algorithm learns by playing against itself without ...
research
07/04/2021

Improve Agents without Retraining: Parallel Tree Search with Off-Policy Correction

Tree Search (TS) is crucial to some of the most influential successes in...
research
05/14/2019

Learning Policies from Self-Play with Policy Gradients and MCTS Value Estimates

In recent years, state-of-the-art game-playing agents often involve poli...
research
11/27/2019

Improving Fictitious Play Reinforcement Learning with Expanding Models

Fictitious play with reinforcement learning is a general and effective f...
research
01/06/2016

Angrier Birds: Bayesian reinforcement learning

We train a reinforcement learner to play a simplified version of the gam...

Please sign up or login with your details

Forgot password? Click here to reset