Variational Monte Carlo simulation with tensor networks of a pure Z3 gauge theory in (2+1)d

by Patrick Emonts (Max-Planck Institute of Quantum Optics & Center for Quantum Science and Technology)

Monday 7 Feb 2022, 14:00 → 16:00 US/Eastern

Gauge theories are at the heart of our understanding of modern physics. Since they are notoriously hard to solve, they are often studied in the form of a lattice regularization as lattice gauge theories (LGT).
Common Monte Carlo algorithms are focused on formulations of LGTs in terms of actions.
This formulation has large computational benefits due to the locality of the action.
However, there are drawbacks as well. The simulation can suffer from the infamous sign problem and the time-evolution of observables is challenging.
In this talk, we will have a look at lattice gauge theories from the Hamiltonian perspective.
We use fermionic gauged Gaussian entangled pair states (fGGPEPS) as an Ansatz in a variational Monte Carlo procedure.
The states are locally gauge invariant by construction and enable sign-problem free sampling of any LGT Hamiltonian.
We demonstrate the algorithm for a pure Z(3) Kogut-Susskind Hamiltonian in (2+1) dimensions.
Details can be found in [Phys. Rev. D 102, 074501 (2020)].