论文标题
实用的大规模平行蒙特卡洛树搜索应用于分子设计
Practical Massively Parallel Monte-Carlo Tree Search Applied to Molecular Design
论文作者
论文摘要
从许多示例中，例如加强学习应用程序，使用大量的计算资源来训练神经网络是普遍的做法。但是，尽管大规模并行计算通常用于训练模型，但很少用于搜索组合优化问题的解决方案。在本文中，我们提出了一种新颖的平行蒙特卡洛树搜索（MP-MCTS）算法，该算法有效地适用于1,000名工人规模，并将其应用于分子设计。这是将分布式MCT应用于现实世界和非游戏问题的第一项工作。大规模并行MCT上的现有工作在最多100名工人的推出数量方面显示出有效的可伸缩性，但遇到了解决方案质量的降解。 MP-MCT将搜索质量保持较大规模，并且通过在256个CPU内核上运行MP-MCT仅10分钟，我们获得了与非平行MCT相似的候选分子，持续42小时。此外，我们基于并行MCT（与简单的RNN模型结合使用）的结果极大地超过了现有的最新工作。我们的方法是通用的，有望加快MCT的其他应用。
It is common practice to use large computational resources to train neural networks, as is known from many examples, such as reinforcement learning applications. However, while massively parallel computing is often used for training models, it is rarely used for searching solutions for combinatorial optimization problems. In this paper, we propose a novel massively parallel Monte-Carlo Tree Search (MP-MCTS) algorithm that works efficiently for 1,000 worker scale, and apply it to molecular design. This is the first work that applies distributed MCTS to a real-world and non-game problem. Existing work on large-scale parallel MCTS show efficient scalability in terms of the number of rollouts up to 100 workers, but suffer from the degradation in the quality of the solutions. MP-MCTS maintains the search quality at larger scale, and by running MP-MCTS on 256 CPU cores for only 10 minutes, we obtained candidate molecules having similar score to non-parallel MCTS running for 42 hours. Moreover, our results based on parallel MCTS (combined with a simple RNN model) significantly outperforms existing state-of-the-art work. Our method is generic and is expected to speed up other applications of MCTS.
