Improving GFlowNets for low data drug discovery

GFlowNets are generative networks that consider the generation process as a flow network. They are theoretically better at handling the difficult cases of sampling different modes of a distribution and when different sampling trajectories can yield the same final state. In this work, we hope to explore various improvements to GFlowNets to make them suitable alternatives for molecular generation and optimization in drug discovery. The research can expand the capabilities of chemists and could enable the discovery of new drugs faster and with less resources.

Faculty Supervisor:

Doina Precup

Student:

Partner:

Valence Discovery Inc

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

McGill University

Program:

Accelerate

Current openings

Find the perfect opportunity to put your academic skills and knowledge into practice!

Government

International

Load More

International

Load More