Probing High-Dimensional Spaces with AI: From Theory Design to Parameter Inference in Particle Physics

by Dr Aishik Ghosh (Georgia Institute of Technology, Atlanta)

Monday 29 Sept 2025, 14:30 → 16:00 Asia/Kolkata

Ag 66 (TIFR, Mumbai)

When confronted with high-dimensional problems, physicists have traditionally relied on intuition to break them down into manageable pieces. But do these simplifications compromise our physics reach? In this talk, we will question core assumptions underlying statistical methods at the foundation of experimental particle physics, such as the likelihood ratio test, as well as mathematical simplifications in theoretical particle physics, aided by some powerful machine learning (ML) tools. We will see that ML allows us to directly probe high-dimensional data at the LHC and achieve a level of precision once thought impossible. These algorithms also let us design theories using mathematical tools for which physicists are yet to build intuition.