Applications of Artificial Intelligence to the Analysis of Chemical and Structural Data

在新功能材料的开发中，结构表征数据的分析是一个耗时的阶段。最近实验室自动化的快速发展意味着数百甚至数千个化学实验可以在没有重大人为干预的情况下进行。然而，确定这些反应的结果可能是具有挑战性的，特别是当固体结构对识别有前途的新产品很重要时。虽然固体表征数据（如x射线衍射）的收集可以集成到自动化工作流程中，但分析数据通常仍然需要人类科学家的大量输入，这阻碍了真正自主实验室工作流程的进展。该博士项目旨在利用人工智能提供一个自动化的过程来分析固态表征数据，最初专注于粉末x射线衍射。使用机器学习方法对数据进行快速分析将是对实验结果提供反馈的关键；这些信息可以决定下一步的实验，因此是实现自主化学实验室的关键组成部分。

The analysis of structural characterisation data is a time-consuming stage in the development of new functional materials. Recent rapid progress in the development of lab automation means that hundreds or even thousands of chemical experiments can be performed without significant human intervention. However, determining the outcome of these reactions can be challenging, particularly when the solid state structure is important in identifying promising new products. While the collection of solid characterisation data, such as X-ray diffraction, can be integrated into an automated workflow, analysing the data often still requires significant input from a human scientist, which hampers progress towards a truly autonomous lab workflow. This PhD project aims to use artificial intelligence to deliver an automated process to analyse solid state characterisation data, initially focusing on powder X-ray diffraction. The use of machine learning methods to perform rapid analyses of data will be key to providing feedback on the outcome of experiments; information that can inform decisions about the next set of experiments, and hence a key component of achieving an autonomous chemistry lab.