化学学报 上一篇    下一篇

综述

人工智能助力当代化学研究

朱博阳a, 吴睿龙a, 于曦a,b   

  1. a 天津大学化学系;
    b 天津市分子光电科学重点实验室 天津 300072
  • 发布日期:2020-08-21
  • 通讯作者: 朱博阳, 于曦 E-mail:luciszhu@outlook.com;xi.yu@tju.edu.cn
  • 基金资助:
    该项目得到了国家自然科学基金(21973069,21773169,21872103)、国家重点研究开发计划(2017YFA0204503,2016YFB0401100)、天津大学北洋青年学者计划(2018XRX-0007)和天津大学大学生创新创业训练计划(201910056451)的支持

Artificial Intelligence for Contemporary Chemistry Research

Zhu Boyanga, Wu Ruilonga, Yu Xia,b   

  1. a Department of Chemistry;
    b Tianjin Key Laboratory of Molecular Optoelectronic Sciences, Tianjin University, Tianjin 30072, China
  • Published:2020-08-21

以机器学习为代表的人工智能在当代的科学研究中正在发挥越来越重要的作用。不同于传统的计算机程序,机器学习人工智能可以通过对大量数据的反复分析和自身模型的优化,即“学习”过程,从而在大量的数据中寻找客观事物的相互联系,形成具有更好预测和决策能力的新模型,做出合理的判断。化学研究的特点恰恰是机器学习人工智能的强项。化学研究经常要面对十分复杂的物质体系和实验过程,从而很难通过化学物理原理进行精准的分析和判断。人工智能可以挖掘化学实验中产生的海量实验数据的相关性,帮助化学家做出合理分析预测,大大加速化学研发过程。本文介绍了当代人工智能方法及用其解决化学问题基本原理,并通过具体案例展示了人工智能辅助解决不同化学研发问题的方法以及对应的机器学习算法。将人工智能运用在化学科学的尝试正处于蓬勃上升期,人工智能已经初步展示出对化学研究的强大助力,希望本文能帮助更多的国内的化学工作者了解和运用这一有力的工具。

关键词: 人工智能;机器学习, 化学, 分子描述符, 深度学习

Modern artificial intelligence (AI), represented by machine learning, is playing an increasingly important role in contemporary scientific research. Unlike traditional programs, machine learning artificial intelligence can establish new models by its own with prediction and decision-making capabilities through the repeated analysis of large amounts of data and optimization of its own model, i.e. the "learning" process, so as to find the correlation and the internal laws of the data. The characteristic of chemical research is precisely the strength of machine learning artificial intelligence. In chemistry research, chemists often face very complicated material systems and experimental processes, making it difficult in accurately analyzing and making judgment through the principles of chemistry and physics. Artificial intelligence can mine the internal laws and correlations of the massive experimental data generated in chemical experiments, help chemists make reasonable analysis and prediction, and greatly accelerate the chemical research and development process. This article discussed contemporary artificial intelligence, mainly the machine learning methods and the basic principles using AI to solve chemical problems, and demonstrated the methods of artificial intelligence assisted solutions to different chemical research and development problems and corresponding machine learning algorithms through specific cases. Attempts to apply artificial intelligence to chemical science are in an age of vigorous growth. Artificial intelligence has initially shown a strong boost to chemistry research. We hope this article will help more domestic chemists to understand and apply this powerful tool.

Key words: Artificial Intelligence, Chemistry, Machine Learning, Molecular Descriptor, Deep Learning