(156a) Keynote Talk-Bayesian Optimization for Additive Manufacturing of Thermoelectric Materials and Devices

Discovering and manufacturing functional materials with precisely tuned properties is a central goal of material science and engineering; yet materials discovery and additive manufacturing optimization is often slow and expensive. For example, novel solid-state thermoelectric (TE) materials have the potential to improve energy efficiency by converting waste heat into electricity. However, the performance of many state-of-the-art TE materials remains inadequate for adoption beyond niche applications. Current efforts to optimize photonic sintering, an important step in additive manufacturing of TE devices, rely on expert-driven trial-and-error search which is often extremely time-consuming and without the guarantee of improvement.

In this talk, we argue Bayesian optimization (BO) and related machine learning frameworks offer principled approaches to intelligentially recommend optimized experimental conditions by balancing exploitation and exploration. Specifically, we show successfully application of BO to materials composition (ink formulation), printing, and sintering steps in the additive manufacturing of n-type and p-type TE materials. We highlight the role of "systems thinking" to synergistically fuse human intuition and BO machine learning models, ultimately resulting in the synthesis of record-breaking TE materials with under fifty experiments.