The Surprising Effectiveness of Approximate Value Iteration in Self-Play

2026-09-08

AI Summary

研究使用近似值迭代(AVI)在自对弈中解决计算成本高的问题，发现AVI在多个游戏中比AlphaZero更有效且成本更低。

Abstract

Combining search with function approximation has driven major advances in game-playing programs, making self-play algorithms more competitive than ever. Still, the computational overhead of the most popular methods, based on Monte Carlo Tree Search (MCTS), can be substantial. In this work, we investigate whether simpler methods remain competitive in non-trivial, moderately sized games such as Connect Four, Hex(7x7) and synthetic games. We train a minimal self-play implementation of Approximate Value Iteration (AVI) and use ground-truth oracles for exact evaluation. Contrary to expectations, our results demonstrate the surprising effectiveness of AVI: it learns more accurate value functions than those learned by AlphaZero, while its one-step-lookahead greedy policies remain competitive with MCTS-based policies at substantially lower training and inference costs. Preliminary experiments on Othello and Go(9x9) show that AVI trains stably on larger games and learns effective value functions. These findings suggest that the success of MCTS-based methods may have eclipsed simpler approaches that have become increasingly practical with modern deep-learning tools.

Problem

• Approximate Value Iteration
• self-play
• Monte Carlo Tree Search
• game-playing programs

Innovation

• Approximate Value Iteration
• self-play
• value function accuracy
• training cost reduction
• inference cost reduction