Overview
Explore the intersection of machine learning and high-energy physics theory in this conference talk from the 2022 Snowmass Theory Frontier Conference at KITP. Delve into Jesse Thaler's evolving perspective on applying ML techniques to theoretical physics, including the likelihood ratio trick and deep learning for lattice field theory. Discover how machine learning ingredients can be leveraged across various aspects of theoretical physics, from network architectures to optimal transport for collider geometry. Gain insights into the potential applications of ML in formal theory, Standard Model theory, and beyond Standard Model (BSM) theory. Engage with the dialogue between machine learning and theoretical physics communities, and understand the growing importance of ML in advancing the theory frontier of particle physics.
Syllabus
Intro
My (Evolving) Perspective
The Lens of Machine Learning
Eg Likelihood Ratio Trick
Machine Learning Ingredients
What is the Machine Learning
Deep Learning for Lattice Field Theory (TF05)
Theoretical Priors Network Architectures
Optimal Transport for Collider Geometry
Opening a Dialogue Between Communities
ML for Formal Theory?
ML for SM Theory?
Machine Learning for the Theory Frontier
ML for BSM Theory!
Taught by
Kavli Institute for Theoretical Physics