2018/03/31 by Anjana Talapatra, S. Boluki, Shahin Boluki +8 · 73 citations
Engineering · Materials Science · Physics and Astronomy · #Advanced materials and composites #Bayesian inference #Bayesian optimization #Bayesian probability #Boron and Carbon Nanomaterials Research #Design of experiments #Framing (construction) #Machine Learning in Materials Science #Space (punctuation) #Ternary operation #cond-mat.mtrl-sci #physics.comp-ph
paper · pdf · doi:10.1103/physrevmaterials.2.113803
published in Physical Review Materials 2(11) (American Physical Society)
arxiv created 2018/10/30 · openalex created_date 2018/11/09 · openalex publication_date 2018/11/26 · arxiv updated 2018/12/05 · openalex updated_date 2026/08/06
The accelerated exploration of the materials space in order to identify configurations with optimal properties is an ongoing challenge. Current paradigms are typically centered around the idea of performing this exploration through high-throughput experimentation/computation. Such approaches, however, do not account for---the always present---constraints in resources available. Recently this problem has been addressed by framing materials discovery as an optimal experiment design. This work augments earlier efforts by putting forward a framework that efficiently explores the materials design space not only accounting for resource constraints but also incorporating the notion of model uncertainty. The resulting approach combines Bayesian model averaging within Bayesian optimization in order to realize a system capable of autonomously and adaptively learning not only the most promising regions in the materials space but also the models that most efficiently guide such exploration. The framework is demonstrated by efficiently exploring the MAX ternary carbide/nitride space through density functional theory (DFT) calculations.