HEP-JEPA
Pretrains by predicting the hidden parts of a jet in an abstract representation instead of reconstructing the particles themselves.
- Made by
- Jai Bardhan and others
- Field
- Collider physics
- Learns from
- JetClass, 100 million jets
- Size
- Not stated
- Open weights
- Not stated
- First public
- Feb 2025
- Source
- arxiv.org/abs/2502.03933
How it relates to Flare
Shows there is more than one way to pretrain on collisions. CERN has not said which approach Flare will take.
Links
Every figure here is from the source above, checked 17 October 2026. Something changed? Email [email protected].