Jan 15 – 17, 2020
Kimmel Center for University Life
America/New_York timezone

Variational Autoencoders for Anomalous Jet Tagging

Jan 16, 2020, 3:30 PM
KC 802 (Kimmel Center for University Life)

KC 802

Kimmel Center for University Life

60 Washington Square S, New York, NY 10012


Taoli Cheng (University of Montreal)


We present a detailed study on Variational Autoencoders (VAEs) performing in anomalous jet tagging. By taking in low-level jet constituents' information, and only training with background jets in an unsupervised manner, the VAE is able to encode important information for reconstructing jets, while learning an expressive posterior distribution in the latent space. The encoder (inference) and decoder (generation) can be used together or seperately to identify out-of-distribution anomalous jets. We employed different techniques to regularize the latent representation, and show how the behavior changes. When using VAE as anomaly detector, we present two approaches to detect anomalies: directly comparing in input space or, instead, working in latent space. Results of tagging performance for different jet types and full kinematic range are shown. In addition, we also study a few tricks to make VAE more sensitive to anomalies.

Primary author

Taoli Cheng (University of Montreal)


Jean-Francois Arguin (Universite de Montreal (CA)) Tobias Golling (Universite de Geneve (CH)) Johnny Raine (Universite de Geneve (CH)) Takuya Nobe (University of Tokyo (JP)) Jacinthe Pilette (Montreal) Julien Leissner-Martin (University of Montreal) Amir Farbin (University of Texas at Arlington (US)) Mr Debottam Bakshi Gupta (University of Texas at Arlington (US))

Presentation materials