Jul 6 – 8, 2021
Europe/Zurich timezone

Better latent spaces for better autoencoders

Jul 8, 2021, 11:00 AM


Dr Barry Dillon (University of Heidelberg)


Autoencoders as tools behind anomaly searches at the LHC have the structural problem that they only work in one direction, extracting jets with higher complexity but not the other way around. To address this, we derive classifiers from the latent space of (variational) autoencoders, specifically in Gaussian mixture and Dirichlet latent spaces. In particular, the Dirichlet setup solves the problem and improves both the performance and the interpretability of the networks.

Academic Rank Postdoctoral researcher
Affiliation University of Heidelberg

Primary author

Dr Barry Dillon (University of Heidelberg)

Presentation materials