Deep Graph anomaly detection with contrastive learning for new physics searches


In the search for new physics at the Large Hadron Collider (LHC) a possible approach is to employ anomaly detection techniques to spot events that deviate from the standard model in an unsupervised manner. There have been many such studies using e.g. convolutional autoencoders. In previous GSoC projects, the usage of graph based models have been very successful in generative tasks. In this project we therefore want to employ a graph based architecture to perform anomaly detection on particle collision data.

The intended model is supposed to perform anomaly detection on a graph-level, corresponding literature can be found in the ‘Links’ section.


Total project length: 175/350 hours.

Difficulty Level

Task ideas

Expected results


Please use this link to access the test for this project.



Please DO NOT contact mentors directly by email. Instead, please email with Project Title and include your CV and test results. The mentors will then get in touch with you.

Corresponding Project

Participating Organizations