Closed-loop discovery of out-of-distribution processing protocols by evolutionary search and uncertainty-aware learning
Researchers have developed a novel closed-loop workflow that combines evolutionary search with uncertainty-aware deep kernel learning to discover new processing protocols for materials. This method was applied to ferroelectric thin films, successfully identifying waveform families that enhance nonlinear electromechanical response by activating specific domain-wall segments. The framework is presented as a generalizable approach for out-of-distribution discovery in various scientific and engineering domains, including synthesis, annealing, and battery formation. AI
IMPACT This research demonstrates a novel AI-driven approach for accelerating materials science discovery, potentially reducing experimental costs and time.