Discovering faster matrix multiplication algorithms with reinforcement learning

作者：Alhussein Fawzi, Matej Balog, Aja Huang, Thomas Hubert, Bernardino Romera‐Paredes, Mohammadamin Barekatain, Alexander Andreevich Novikov, Francisco J. R. Ruiz, Julian Schrittwieser, Grzegorz Świrszcz, David Lawrence Silver, Demis Hassabis, Pushmeet Kohli · 发表于：Nature · 年份：2022 · DOI：10.1038/s41586-022-05172-4 · 被引用次数：475 · 研究领域：Tensor decomposition and applications、Evolutionary Algorithms and Applications、Cellular Automata and Applications

Abstract Improving the efficiency of algorithms for fundamental computations can have a widespread impact, as it can affect the overall speed of a large amount of computations. Matrix multiplication is one such primitive task, occurring in many systems—from neural networks to scientific computing routines. The automatic discovery of algorithms using machine learning offers the prospect of reaching beyond human intuition and outperforming the current best human-designed algorithms. However, automating the algorithm discovery procedure is intricate, as the space of possible algorithms is enormous. Here we report a deep reinforcement learning approach based on AlphaZero 1 for discovering efficient and provably correct algorithms for the multiplication of arbitrary matrices. Our agent, AlphaTensor, is trained to play a single-player game where the objective is finding tensor decompositions within a finite factor space. AlphaTensor discovered algorithms that outperform the state-of-the-art complexity for many matrix sizes. Particularly relevant is the case of 4 × 4 matrices in a finite field, where AlphaTensor’s algorithm improves on Strassen’s two-level algorithm for the first time, to our knowledge, since its discovery 50 years ago 2 . We further showcase the flexibility of AlphaTensor through different use-cases: algorithms with state-of-the-art complexity for structured matrix multiplication and improved practical efficiency by optimizing matrix multiplication for runtime on s...