Science at the Edge Friday October 12th,2018 1400 Biomedical Physical Science 11:30 am (Refreshments 11:15) Speaker: Peter Frazier Bayesian Optimization for Materials Design and Drug Discovery The search for new materials and medicines often relies on burdensome experimental evaluation of large numbers of material or drug candidates. Using machine learning to predict biological or chemical properties in silico before experimental evaluation has drawn immense interest as a way to reduce this burden. In this talk, we argue that while machine learning is promising, it is not magic. Its predictions are often incorrect, and its successful use in materials and drug discovery often requires using it in a way that is robust to these prediction errors. We show how ideas from Bayesian optimization can be used to build in a specific kind of robustness to experiments informed by a machine learning predictive model. This approach can dramatically improve the effectiveness of machine learning in materials discovery tasks. We show how these ideas were used in a recent collaboration to discover short peptides with specific enzymatic activity, which led to the creation of an orthogonal peptide labeling system. This is joint work with L. Tallorin, J. Wang, W.E. Kim, S. Sahua, N.M. Kosa, P. Yang, M. Thompson, M.K. Gilson, N.C. Gianneschi and M. D. Burkart Lerena R. Heintzelman Department of Physics & Astronomy Michigan State University 567 Wilson Rd. Room 3261 East Lansing, MI 48824 517-884-5513