(This project is co-supervised with David Buterez, and maybe Pietro Lio)

We have previously done a gradient-based Network Architecture Search algorithm for Graph Neural Networks (GNNs). High-throughput screening (HTS) is one leading technique for hit identification in drug discovery. David’s prior work on this has demonstrated that GNN models can potentially help in this process:

Multi-fidelity machine learning models for improved high-throughput screening predictions

The project will focus on developing Automated ML methods for this particular application.