A GitLab upgrade is scheduled for Monday, June 30, 2025. The service will be unavailable for a few minutes in the morning. We'll keep you posted on the progress of the upgrade on the Mattermost channel: https://mattermost.inria.fr/devel/channels/gitlab. We recommend that you do not work on the platform until an announcement indicates that maintenance is complete.
I present a method for computing statistically guaranteed confidence bands for functional surrogate modes: surrogate models which map between function spaces, motivated by the need build reliable physics emulators. The method constructs nested confidence sets on a low-dimensional representation (an SVD) of the surrogate model’s prediction error, and then maps these sets to the prediction space using set-propagation techniques. The result is conformal-like coverage guaranteed prediction sets for functional surrogate models. We use zonotopes as basis of the set construction, due to their well-studied set-propagation and verification properties. The method is model agnostic and can thus be applied to complex Sci-ML models, including Neural Operators, but also in simpler settings. An important step is a technique to capture the truncation error of the SVD, ensuring the guarantees of the method.
I present a method for computing statistically guaranteed confidence bands for functional surrogate modes: surrogate models which map between function spaces, motivated by the need build reliable physics emulators. The method constructs nested confidence sets on a low-dimensional representation (an SVD) of the surrogate model’s prediction error, and then maps these sets to the prediction space using set-propagation techniques. The result is conformal-like coverage guaranteed prediction sets for functional surrogate models. We use zonotopes as basis of the set construction, due to their well-studied set-propagation and verification properties. The method is model agnostic and can thus be applied to complex Sci-ML models, including Neural Operators, but also in simpler settings. An important step is a technique to capture the truncation error of the SVD, ensuring the guarantees of the method.
A preprint is available here: https://doi.org/10.48550/arXiv.2501.18426
A preprint is available here: [arXiv:arXiv.2501.18426](https://doi.org/10.48550/arXiv.2501.18426)
