Active machine learning for novel molecule discovery
Large-scale screening of new molecules is expensive and time-consuming. In the first part of the talk, I will introduce a novel active machine learning-based framework that statistically minimizes the number of wet-lab experiments needed to design new Antimicrobial Peptides (AMPs), while ensuring a high diversity and novelty of generated AMPs sequences, in multi-rounds of wet-lab AMP screening settings. To enhance the active learning, a powerful and robust uncertainty estimation method is needed. In the second part of the talk, a novel generative flow network-based dropout method called “GFlowOut” will be introduced and how it can help uncertainty estimation and active learning will be explained.
Dianbo Liu is a postdoctoral machine learning researcher in the group of Prof. Yoshua Bengio (Turing Award 2018) at the Mila-Quebec AI institute. He additionally leads the humanitarian AI team of 19 researchers. Dianbo’s research spans both fundamental machine learning and its applications in biomedical informatics. Prior to joining the Bengio team, Dianbo worked and studied at the University of Dundee, Harvard University and Massachusetts Institute of Technology. Dianbo co-found and served as the first CTO of Secure AI Labs, MA, USA, which is a MIT spin-off that focuses on federated learning for social good. In his personal life, Dianbo is a stand-up comedian in training.