Predictive Fidelity, Interpretability, and Resilience of Machine Learning Methods Applied to Scientific Simulations

Whitepaper submitted to the 2017 DOE ASCR Applied Math Meeting<div><br></div><div>All authors are affiliated with Sandia National Laboratories<br><div><br></div><div>From the list at https://www.orau.gov/ascr-appliedmath-pi2017/whitepaper-questions.htm, this white paper addresses the questions 2b and 4b.</div><div><br></div><div><br></div>