Minisymposium 207, Operator Learning in Mechanics with Robust Error Estimation, Control, and Generalization, organized with Danial Faghihi and Serge Prudhomme, ran across two sessions on 23 June at the 20th U.S. National Congress on Theoretical and Applied Mechanics in Pasadena, California. The two sessions included eight presentations on Bayesian neural operators, residual-based error correction, reinforcement learning, and generalization in mechanics.
The organizing question was the one this group keeps returning to: not whether a learned operator is accurate on a test set, but whether it can say how wrong it is on the input in front of it, and what happens when that input is unlike anything it was trained on.
Thanks to the speakers and to everyone who came, and for the conversation that carried on over dinner afterwards.
