News Archives

RSS Feed

[Colloquium] Predictions and Machine Learning: A personal viewpoint

October 19, 2007

Watch Colloquium: 


  • Date: Friday, October 19th, 2007 
  • Time: 1 pm — 2:30 pm 
  • Place: ME 218

Tudor I. Oprea
Professor of Biochemistry and Molecular Biology and Chief of the Division of Biocomputing
UNM School of Medicine

Abstract: From the perspective of molecular bioactivity, machine learning is a widely used tool that streamlines the R&D process and enables unexpected discoveries. This talk will first showcase the use of machine learning in small molecule discovery, at UNM Biocomputing. Then we discuss the impact of the “unknown unknown” in our ability to perform accurate predictions, which is inherently limited by the past. The Black Swan, a metaphor for highly improbable, high impact events, warns us to understand the limits of predictions and expect the “unknown unknowns” everywhere, from war to science and industry. We further argue that the structure of the pharmaceutical and flavor industries are driven by, and impacted by both positive and negative Black Swans.

Bio: Tudor I. Oprea has an MD and a PhD from the Universityof Medicine and Pharmacy in Timisoara, Romania. His past appointments were with the University of Utrecht, the Netherlands, Washington University in St.Louis, MO, and Los Alamos National Lab, before spending 6 years at AstraZeneca, Sweden. Since 2002, Tudor is Professor of Biochemistry and Molecular Biology, and Chief of the Division of Biocomputing at the UNM School of Medicine. Tudor works in the areas of QSAR, lead identification and optimization, virtual screening and drug discovery. He has co-authored two books, seventeen book chapters, more than sixty peer-reviewed papers and several patent disclosures. He received the 2002 Hansch Award from the QSAR and Molecular Modeling Society and serves as the Chair of the Cheminformatics and QSAR Society. He is the founder and CEO of Sunset Molecular Discovery LLC.