Research talk: Leveraging uncertainty in machine learning to bridge computation and experimentation
19Speaker: Ava Soleimany, Sr. Researcher, Microsoft Health Futures
While machine learning (ML) is poised to have a transformative impact on biomedicine, its success is contingent on experimental efforts that generate relevant datasets and validate model predictions. Critically, such experimentation often requires significant time, personnel, material, and monetary investments. Computational methods to inform experimental modeling could help alleviate this resource burden and bridge the gap between computational predictions and experimental validation. In this talk, we’ll discuss how intelligent ML algorithms that quantify prediction uncertainties could meet this critical need. Using molecular property prediction and drug discovery as a motivating use case, we’ll present a new method for calibrated uncertainty quantification in neural networks and demonstrate its potential to (1) improve sample efficiency via uncertainty-guided active learning and (2) inform experimental validation via targeted virtual screening. More broadly, this work provides generalizable frameworks for how prediction uncertainty can accelerate and guide key steps in experimental lifecycles, opening the door for sustained, iterative feedback between computation and experimentation.
Learn more about the 2021 Microsoft Research Summit: https://Aka.ms/researchsummit