Portrait of Pascal Germain

Pascal Germain

Associate Academic Member
Canada CIFAR AI Chair
Assistant Professor, Université Laval, Department of Computer Science and Software Engineering
Research Topics
Machine Learning Theory
Representation Learning
Website
Google Scholar
LinkedIn
X

Biography

Pascal Germain is an assistant professor in the Computer Science and Software Engineering Department at Université Laval, and a Canada CIFAR AI Chair. He received his PhD in computer science from Université Laval in 2015 under the supervision of François Laviolette and Mario Marchand. He then pursued his research at the INRIA in France for four years, first as a postdoctoral fellow with Francis Bach’s SIERRA project team, then as a research fellow in the MODAL project team. Germain was also an affiliate member and lecturer in the Mathematics Department at the University of Lille. He returned to his alma mater, Université Laval, in 2019 to take on the role of assistant professor.

He is currently a member of Université Laval’s Big Data Research Centre (CRDM) and Institute Intelligence and Data (IID), as well as an associate academic member of Mila – Quebec Artificial Intelligence Institute. His research interests include machine learning, in particular statistical machine learning theory, transfer learning and interpretable predictor learning.

Current Students

Mathieu Bazinet
PhD - Université Laval
Github
Thibaud Godon
PhD - Université Laval
Co-supervisor :
Alexandre Drouin
Github
Google Scholar
Benjamin Leblanc
PhD - Université Laval
Website
Google Scholar
Diarra Mbacke
PhD - Université Laval
Google Scholar

Publications

Sample compression unleashed : New generalization bounds for real valued losses
Mathieu Bazinet
Valentina Zantedeschi
Pascal Germain
2024-10-10
NeurIPS.cc/2024/Workshop/M3L (poster)
openreview.net
openreview.net
Seeking Interpretability and Explainability in Binary Activated Neural Networks
Benjamin Leblanc
Pascal Germain
2024-07-10
Communications in Computer and Information Science (published)
doi.org
arxiv.org
Phoneme Discretized Saliency Maps for Explainable Detection of AI-Generated Voice
Shubham Gupta
Mirco Ravanelli
Pascal Germain
Cem (Yusuf) Subakan
2024-06-14
ArXiv (preprint)
doi.org
arxiv.org
A general framework for the practical disintegration of PAC-Bayesian bounds
Paul Viallard
Pascal Germain
Amaury Habrard
Emilie Morvant
2023-10-11
Machine-mediated learning (published)
doi.org
arxiv.org
Statistical Guarantees for Variational Autoencoders using PAC-Bayesian Theory
Sokhna Diarra Mbacke
Florence Clerc
Pascal Germain
openreview.net
Statistical Guarantees for Variational Autoencoders using PAC-Bayesian Theory
Sokhna Diarra Mbacke
Florence Clerc
Pascal Germain
2023-09-21
NeurIPS.cc/2023/Conference (spotlight)
openreview.net
openreview.net
PAC-Bayesian Generalization Bounds for Adversarial Generative Models
Sokhna Diarra Mbacke
Florence Clerc
Pascal Germain
2023-07-03
Proceedings of the 40th International Conference on Machine Learning (published)
doi.org
openreview.net
Invariant Causal Set Covering Machines
Thibaud Godon
Baptiste Bauvin
Pascal Germain
Jacques Corbeil
Alexandre Drouin
2023-06-07
ArXiv (preprint)
doi.org
arxiv.org
PAC-Bayesian Learning of Aggregated Binary Activated Neural Networks with Probabilities over Representations
Louis Fortier-Dubois
Gaël Letarte
Benjamin Leblanc
François Laviolette
Pascal Germain
2023-06-05
Proceedings of the Canadian Conference on Artificial Intelligence (published)
doi.org
arxiv.org
PAC-Bayesian Generalization Bounds for Adversarial Generative Models
Sokhna Diarra Mbacke
Florence Clerc
Pascal Germain
We extend PAC-Bayesian theory to generative models and develop generalization bounds for models based on the Wasserstein distance and the to… (see more)tal variation distance. Our first result on the Wasserstein distance assumes the instance space is bounded, while our second result takes advantage of dimensionality reduction. Our results naturally apply to Wasserstein GANs and Energy-Based GANs, and our bounds provide new training objectives for these two. Although our work is mainly theoretical, we perform numerical experiments showing non-vacuous generalization bounds for Wasserstein GANs on synthetic datasets.
2023-01-01
ICML (published)
doi.org
openreview.net
Sample Boosting Algorithm (SamBA) - An interpretable greedy ensemble classifier based on local expertise for fat data
Baptiste Bauvin
Cécile Capponi
Florence Clerc
Pascal Germain
Sokol Koço
Jacques Corbeil
2023-01-01
UAI (published)
proceedings.mlr.press
openreview.net
Interpretable domain adaptation using unsupervised feature selection on pre-trained source models
Luxin Zhang
Pascal Germain
Yacine Kessaci
C. Biernacki
2022-09-01
Neurocomputing (published)
doi.org