Machine Learning for Materials Design

Friday, April 22, 2016 - 1:25pm - 2:25pm
Lind 305
Adama Tandia (Corning Incorporated)
Materials play a very critical role in many aspects of human life and activities. Design of new materials with targeted properties had become very tedious because of the high complexity of the interplay between the components and the more and more increasing customer expectations. Modern glassy materials are no exception. Through a combination of different empirical predictive modeling routes from Machine Learning, atomistic scale modeling, and the existence of large historical data, it has become possible to design materials with optimized properties at a much faster space. Illustration will be based on use of molecular dynamics to compute the glass lattice dilation parameter, and machine learning to investigate glass properties model development.

Adama Tandia holds a Ph.D. degree in Applied Math/Applied Physics from Paul Sabatier University (France), and a Master Degree in the same field at University Paris XII Creteil (France). He worked as a Research Associate at Northwestern University (USA) in the department of Applied Math from February 1998 to November 2000 as a key member of a novel project for crystal growth using Level Set and Monte Carlo methods. He joined Corning Incorporated in December 2000 and led the optical fiber micro-bending modeling effort until June 2002. He then transferred to the department of Modeling & Simulation. Adama is currently involved in process optimization, materials modeling and design, and machine learning.