Research interests
- Gaussian process modeling and time-series
- Uncertainty quantification for computer experiments
- Bayesian statistics
Articles
- B. Kerleguer, C. Cannamela, & J. Garnier (2024), A Bayesian neural network approach to Multi-fidelity surrogate modeling, International Journal for Uncertainty Quantification.
- B. Kerleguer (2023), Multi-Fidelity Surrogate Modeling for Time-Series Outputs, SIAM/ASA Journal on Uncertainty Quantification.
Thesis
- Baptiste Kerleguer. Multi-fidelity surrogate modeling adapted to functional outputs for uncertainty quantification of complex models. Statistics [math.ST]. Institut Polytechnique de Paris, 2022. ⟨tel-04106672⟩.
The jury was chaired by Béatrice Laurent, the thesis was reported by Béatrice Laurent and Serge Guillas. The jury was made up of Josselin Garnier, Serge Guillas, Erwan Scornet, Merlin Keller, Anthony Nouy and Claire Cannamela.
Talks
- JDS 2023 of SFdS, Prise en compte de contraintes physiques dans la métamodélisation pour la simulation numérique
- ONERA remote seminar on uncertainties quantification Multi-fidelity surrogate modeling for uncertainty quantification
- JDS 2022 of SFdS, Méta-modélisation multi-fidélité par processus Gaussien en utilisant une base d'ondelettes
- SIAM UQ contributed talk Multi-Fidelity Gaussian Process Regression for High-Dimensional Code Outputs Using Wavelet Transform Covariance
- MIA seminar Multi-fidelity surrogate modeling for time-series output
- PhD talk at ETICS 2021 Multi-fidelity surrogate model combining Gaussian process regression and Bayesian neural network
- UNCECOMP 2021, Multi-Fidelity surrogate modeling for time-series outputs.
- JDS 2021 of SFdS, Meta-modélisation multi-fidélité combinant processus Gaussiens et réseau de neurones bayésien and Métamodélisation multi-fidélité avec des séries-temporelles en sortie
- Numerical analysis Summer school 2021, Multi-fidelity approaches with Claire Cannamela, slides.
- Talk at ETICS 2020 Multi-fidelity modeling for time-series output
Posters
- Live poster at MASCOT-NUM 2021 PhD day Large dimension multi-fidelity surrogate model : Co-Kriging models compare to neural network approaches
- Poster at MASCOT-NUM 2020 PhD day Multi-fidelity modeling for time-series output