Embracing the Unexpected: AI Methods for Surprise Discoveries in Particle Physics

by Oz Amram (Fermilab)

Tuesday 20 Jan 2026, 11:00 → 12:00 America/Los_Angeles

48/1-112A/B - Redwood A/B (SLAC)

Fundamental physics stands at a crossroads: without clear targets for the next breakthrough, we must cast our net wide and let the data guide us. However, modern experiments produce such vast quantities of highly complex information that traditional analysis methods rely on layer upon layer of assumptions to tame it. This greatly limits our ability to discover things we haven't anticipated. Modern AI methods, which are naturally equipped to work with data in its full native complexity, can offer a new path forward. In this talk I will present how methods such as anomaly detection, generative models and foundation models are being deployed in particle physics and discuss how they may rekindle the possibility of genuine surprise.

 

https://stanford.zoom.us/j/98973156241?pwd=cEU5RFdlVXoyc0JTeTlDMkozKzQ5UT09

 

 

SLAC_AI_Amram.pdf