ENBIS & JMP Short Course: Bayesian Desirability Functions – A Modern Approach to Efficient Product and Process Design
In industrial R&D and process improvement we experiment to understand complex, multifactor systems. There has been tremendous interest in applying "Bayesian Optimization" or "Active Learning" to efficiently innovate new products and processes. These approaches to sequential experimentation promise greater speed while being more approachable to scientists and engineers. Generalizing Bayesian Optimization (BayesOpt) to real world complex problems involving multiple responses has proven challenging because in its standard formulation the BayesOpt approach is inherently limited to a single response. In this webinar we review the basics of Gaussian Process regression modeling and the standard approach to BayesOpt. We then introduce the generalization to multiple responses via the Bayesian Desirability framework. We will demonstrate the efficiency and approachability of the technique using new capabilities in JMP Pro 19.
Materials to try these methods yourself in JMP Pro version 19.0 are attached. If you are a student or academic researcher, you can get free access to JMP Student Edition here.
Chris Gotwalt.jpg
Recording and materials from this online short course, 03 December 2025. Presented by Chris Gotwalt and @Phil_Kay .
Re: ENBIS & JMP Short Course: Bayesian Desirability Functions – A Modern Approach to Efficient Product and Process DesignHi Dr. Kay, I'm a student who built a free, open-source tool that runs formulation experiments through screening, confirmation replicates, and Monte Carlo robustness. I saw your Bayesian desirability post. One question, asking to learn: where would a DOE educator object first in that ladder? Repo: https://github.com/TM289012/reformulation-assurance. Nothing for sale. Thanks, TM
