标题： 迈向智能主动粒子 显示英文标题

标题： Towards Intelligent Active Particles

摘要： 在本书章节中，我们描述了人工智能，特别是机器学习在活性物质系统中的最新应用。 活性物质由能够自我推进的代理或粒子组成。 虽然生物代理如细菌、鱼或鸟自然具有一定程度的“智能”，但合成活性粒子如胶体微泳器和电子机器人可以配备不同水平的人工智能，要么内部（如机器人），要么通过动态外部控制系统。 本章简要讨论了现有方法，使合成粒子越来越“智能”，然后重点介绍了使用机器学习来解决活性粒子的导航和通信问题。 基本问题是如何通过复杂环境最优地引导单个活性代理以达到或发现目标，以及活性粒子需要如何合作以高效地从其复杂环境中收集目标分布（例如营养物或毒素）。 显示英文摘要

摘要： In this book chapter we describe recent applications of artificial intelligence and in particular machine learning to active matter systems. Active matter is composed of agents, or particles, that are capable of propelling themselves. While biological agents like bacteria, fish or birds naturally possess a certain degree of "intelligence", synthetic active particles like colloidal microswimmers and electronic robots can be equipped with different levels of artificial intelligence, either internally (as for robots) or via a dynamic external control system. This book chapter briefly discusses existing approaches to make synthetic particles increasingly "intelligent" and then focuses on the usage of machine learning to approach navigation and communication problems of active particles. Basic questions are how to steer a single active agent through a complex environment to reach or discover a target in an optimal way and how active particles need to cooperate to efficiently collect a distribution of targets (e.g. nutrients or toxins) from their complex environment.