Variational Wavefunctions, Machine Learning Architecture for Fermions and Gauge Theories
Institute for Pure & Applied Mathematics (IPAM) via YouTube
Overview
Explore a conference talk on advanced machine learning techniques for quantum systems, focusing on variational wavefunctions for fermions and gauge theories. Delve into the combination of machine learning architectures with physics-inspired approaches to build symmetry-preserving variational wave-functions. Examine the use of configuration-dependent Slater determinants for fermions and modified autoregressive neural networks for gauge theories. Discover how these methods lead to accurate results across various quantum systems, including larger systems and symmetry restoration. Gain insights into neural network flow, R by R matrix calculations, and slave fermion concepts in quantum mechanics.
Syllabus
Intro
Machine Learning for Fermions
Neural Net Backflow
Neural Network
Generalizations
BDG State
Larger Determinants
Multidetermining Expansion
Results
Neural Net
Larger Systems
Symmetry Restoration
Neural Network Flow
R by R Matrix
Decay Event Measurement Evolution
Slave Fermions
Taught by
Institute for Pure & Applied Mathematics (IPAM)