Agathe Senellart

photo_agathe_senellart.jpg

I am currently a PhD student under the supervision of Stéphanie Allassonnière in the HeKA team located at PariSanté Campus. I work on Variational Autoencoders based methods with applications to medical imaging. During my PhD, I have worked on Multimodal VAEs methods, which led me to develop the python library MultiVae. More recently I have been working on Unsupervised Anomaly Detection methods in brain images, in collaboration with the Aramis team at the Paris Brain Institute.

news

Mar 20, 2026 Our work Mitigating the reconstruction-detection trade-off in VAE-based unsupervised anomaly detection was accepted for an oral presentation at ISBI 2026 in London !
Mar 15, 2026 We will be at OHBM-2026 🧠 in Bordeaux to present our work on latent maximum a posteriori optimisation for anomaly detection !
Mar 11, 2026 We presented a poster at IABM 2026 in Lyon!
Feb 15, 2026 I was attending the workshop AI across Scales at the Newton Institute in Cambridge and presented a poster about our work on unsupervised anomaly detection.
Nov 13, 2025 Our paper “Bridging the inference gap in Multimodal Variational Autoencoders” was accepted for publication in JDSSV !
Jun 05, 2025 Our paper “MultiVae: A Python package for Multimodal Variational Autoencoders on Partial Datasets.” was accepted for publication in JOSS ! The library is available here :sparkles: :smile:
Jun 01, 2025 I will be presenting my work on Multimodal VAEs at the JDS-2025 in Marseille !