Biologically-informed Representations for Integrating Heterogeneous Multi-Modal data for Deep Neural Networks in Drug Discovery
Lead Research Organisation:
University of Cambridge
Department Name: Biochemistry
Abstract
The overall goals of this project are two-fold: producing highly effective models for specific tasks within molecular interaction prediction, as well as methodological improvements for deep learning algorithms on graph-structured data. This will largely be focussed on methodological improvements relating to overcoming existing limitations in models for the purposes of molecular biology and drug design.