Showing 1–2 of 2 results for author: Picard-Weibel, A
-
Bayesian Uncertainty Quantification for Anaerobic Digestion models
Authors:
Antoine Picard-Weibel,
Gabriel Capson-Tojo,
Benjamin Guedj,
Roman Moscoviz
Abstract:
Uncertainty quantification is critical for ensuring adequate predictive power of computational models used in biology. Focusing on two anaerobic digestion models, this article introduces a novel generalized Bayesian procedure, called VarBUQ, ensuring a correct tradeoff between flexibility and computational cost. A benchmark against three existing methods (Fisher's information, bootstrap** and Be…
▽ More
Uncertainty quantification is critical for ensuring adequate predictive power of computational models used in biology. Focusing on two anaerobic digestion models, this article introduces a novel generalized Bayesian procedure, called VarBUQ, ensuring a correct tradeoff between flexibility and computational cost. A benchmark against three existing methods (Fisher's information, bootstrap** and Beale's criteria) was conducted using synthetic data. This Bayesian procedure offered a good compromise between fitting ability and confidence estimation, while the other methods proved to be repeatedly overconfident. The method's performances notably benefitted from inductive bias brought by the prior distribution, although it requires careful construction. This article advocates for more systematic consideration of uncertainty for anaerobic digestion models and showcases a new, computationally efficient Bayesian method. To facilitate future implementations, a Python package called 'aduq' is made available.
△ Less
Submitted 30 May, 2024;
originally announced May 2024.
-
On change of measure inequalities for $f$-divergences
Authors:
Antoine Picard-Weibel,
Benjamin Guedj
Abstract:
We propose new change of measure inequalities based on $f$-divergences (of which the Kullback-Leibler divergence is a particular case). Our strategy relies on combining the Legendre transform of $f$-divergences and the Young-Fenchel inequality. By exploiting these new change of measure inequalities, we derive new PAC-Bayesian generalisation bounds with a complexity involving $f$-divergences, and h…
▽ More
We propose new change of measure inequalities based on $f$-divergences (of which the Kullback-Leibler divergence is a particular case). Our strategy relies on combining the Legendre transform of $f$-divergences and the Young-Fenchel inequality. By exploiting these new change of measure inequalities, we derive new PAC-Bayesian generalisation bounds with a complexity involving $f$-divergences, and holding in mostly unchartered settings (such as heavy-tailed losses). We instantiate our results for the most popular $f$-divergences.
△ Less
Submitted 11 February, 2022;
originally announced February 2022.
