When Bayesian neural networks remember too much
Published:
A Bayesian posterior that accumulates every observation ends up too certain to learn anything new. One forgetting term gives the variance a floor, and the model a memory window.
Posts by Tags
Bayesian
Active Continual Learning with Metaplastic Binary Bayesian Neural Networks
Published:
We are excited to share that our paper, Active Continual Learning with Metaplastic Binary Bayesian Neural Networks, has been accepted as a poster at ICML 2026 in Seoul, South Korea!
A geometrical interpretation of variational methods in machine learning
Published:
Variational methods are often introduced through information theory: learning means finding an approximate distribution that remains close to a target Bayesian posterior. This has a geometrical interpretation.
Bayesian Continual Learning in Nature Communications
Published:
I’m super happy to announce that Djohan Bonnet, Damien Querlioz and I, as well as co-authors, recently published a paper about Bayesian continual learning and forgetting in neural networks in Nature Communications.
Binary neural networks
When Bayesian neural networks remember too much
Published:
A Bayesian posterior that accumulates every observation ends up too certain to learn anything new. One forgetting term gives the variance a floor, and the model a memory window.
Continual learning
When Bayesian neural networks remember too much
Published:
A Bayesian posterior that accumulates every observation ends up too certain to learn anything new. One forgetting term gives the variance a floor, and the model a memory window.
Active Continual Learning with Metaplastic Binary Bayesian Neural Networks
Published:
We are excited to share that our paper, Active Continual Learning with Metaplastic Binary Bayesian Neural Networks, has been accepted as a poster at ICML 2026 in Seoul, South Korea!
Bayesian Continual Learning in Nature Communications
Published:
I’m super happy to announce that Djohan Bonnet, Damien Querlioz and I, as well as co-authors, recently published a paper about Bayesian continual learning and forgetting in neural networks in Nature Communications.
Geometry
A geometrical interpretation of variational methods in machine learning
Published:
Variational methods are often introduced through information theory: learning means finding an approximate distribution that remains close to a target Bayesian posterior. This has a geometrical interpretation.
Interview
Itinérances - Ph.D Student Portrait - Lara Couronné
Published:
This portrait celebrates Lara Couronné, a brilliant scientist whose journey blends determination, passion, and creativity.
Itinérances - Ph.D Student Portrait - Kellian Cottart
Published:
Itinérances is a portrait of my journey as a Ph.D. student, at the crossroads of scientific research, personal passion, and intellectual curiosity.
Machine Learning
When Bayesian neural networks remember too much
Published:
A Bayesian posterior that accumulates every observation ends up too certain to learn anything new. One forgetting term gives the variance a floor, and the model a memory window.
Active Continual Learning with Metaplastic Binary Bayesian Neural Networks
Published:
We are excited to share that our paper, Active Continual Learning with Metaplastic Binary Bayesian Neural Networks, has been accepted as a poster at ICML 2026 in Seoul, South Korea!
A geometrical interpretation of variational methods in machine learning
Published:
Variational methods are often introduced through information theory: learning means finding an approximate distribution that remains close to a target Bayesian posterior. This has a geometrical interpretation.
Bayesian Continual Learning in Nature Communications
Published:
I’m super happy to announce that Djohan Bonnet, Damien Querlioz and I, as well as co-authors, recently published a paper about Bayesian continual learning and forgetting in neural networks in Nature Communications.
Itinérances - Ph.D Student Portrait - Kellian Cottart
Published:
Itinérances is a portrait of my journey as a Ph.D. student, at the crossroads of scientific research, personal passion, and intellectual curiosity.
Personal
Itinérances - Ph.D Student Portrait - Lara Couronné
Published:
This portrait celebrates Lara Couronné, a brilliant scientist whose journey blends determination, passion, and creativity.
Itinérances - Ph.D Student Portrait - Kellian Cottart
Published:
Itinérances is a portrait of my journey as a Ph.D. student, at the crossroads of scientific research, personal passion, and intellectual curiosity.
Quantum Physics
Itinérances - Ph.D Student Portrait - Lara Couronné
Published:
This portrait celebrates Lara Couronné, a brilliant scientist whose journey blends determination, passion, and creativity.
Release
Active Continual Learning with Metaplastic Binary Bayesian Neural Networks
Published:
We are excited to share that our paper, Active Continual Learning with Metaplastic Binary Bayesian Neural Networks, has been accepted as a poster at ICML 2026 in Seoul, South Korea!
Bayesian Continual Learning in Nature Communications
Published:
I’m super happy to announce that Djohan Bonnet, Damien Querlioz and I, as well as co-authors, recently published a paper about Bayesian continual learning and forgetting in neural networks in Nature Communications.
Variational Inference
When Bayesian neural networks remember too much
Published:
A Bayesian posterior that accumulates every observation ends up too certain to learn anything new. One forgetting term gives the variance a floor, and the model a memory window.
Active Continual Learning with Metaplastic Binary Bayesian Neural Networks
Published:
We are excited to share that our paper, Active Continual Learning with Metaplastic Binary Bayesian Neural Networks, has been accepted as a poster at ICML 2026 in Seoul, South Korea!
A geometrical interpretation of variational methods in machine learning
Published:
Variational methods are often introduced through information theory: learning means finding an approximate distribution that remains close to a target Bayesian posterior. This has a geometrical interpretation.
Bayesian Continual Learning in Nature Communications
Published:
I’m super happy to announce that Djohan Bonnet, Damien Querlioz and I, as well as co-authors, recently published a paper about Bayesian continual learning and forgetting in neural networks in Nature Communications.