MACHINE LEARNING WEBINAR SERIES
Bayesian Frameworks for Accelerated Alloy Discovery
Speaker: Raymundo Arróyave
Tuesday January 28, 2025 at 2-3 PM CET/8-9 AM EST
This webinar will include a discussion of novel approaches to alloy discovery through Bayesian approaches. The framework that will be presented combines experiments and simulations to arrive at optimal discovery sequences capable of identifying optimal materials with minimal resource utilization. A key element in this approach to iterative materials optimization is the use of CALPHAD-based simulation workflows (executed through Thermo-Calc’s Python API) to calculate phase stability and performance-relevant properties. By using these models, efforts can greatly reduce the space to explore such that the Bayesian exploration is carried out in a much more manageable design space. In the webinar, the speaker will present a few examples where these methods have been used to optimize alloys under multiple objectives and constraints.