Investigation of Molecular Representations, Chemical Databases and Machine Learning Models for Molecular Property Prediction

第一年：一般技能培训和实验室轮换第2-4年：最初将考虑分子性质预测中的机器学习（ML），重点是化学毒理学。接下来将考虑新的分子表示，以确定可能适用于ML模型的任何新表示。最后对目前已应用于分子性质预测的可解释人工智能技术进行了综述。 项目工作将包括三个不同的方面：熟悉和调查数据库的重复搜索，调查新的分子表征，以及应用反事实预测化学毒理学。

Year 1: General skills training and lab rotationsYears 2-4: Initially machine learning (ML) within molecular property prediction will be considered, with a focus on chemical toxicology. Following on new molecular representations will be considered to identify any new representations which may be appropriate for ML models. Finally current explainable AI techniques which have been implemented into molecular property prediction will be reviewed. Project work will include three distinct strands: familiarisation and investigation of databases by duplicate searching, investigation of new molecular representations, and the application of counterfactuals for predicting chemical toxicology.