Group-Equivariant Diffusion Models for Lattice Field Theory

Abstract

Near the critical point, Markov Chain Monte Carlo (MCMC) simulations of lattice quantum field theories (LQFT) become increasingly inefficient due to critical slowing down. In this work, we investigate score-based symmetry-preserving diffusion models as an alternative strategy to sample two-dimensional ${\varphi }^4$ and $\mathrm{U}(1)$ lattice field theories. We develop score networks that are equivariant to a range of group transformations, including global ${\mathbb{Z}}_2$ reflections, local $\mathrm{U}(1)$ rotations, and periodic translations $\mathbb{T}$. The score networks are trained using an augmented training scheme, which significantly improves sample quality in the simulated field theories. We also demonstrate empirically that our symmetry-aware models outperform generic score networks in sample quality, expressivity, and effective sample size.

Updated to match published version in JHEP. 45 pages, 12 figures. Code available at https://gitlab.com/ovega141/diffusion_for_lqft

References in corpus (25)

• SciPy 1.0--Fundamental Algorithms for Scientific Computing in Python
• Array Programming with NumPy
• Diffusion Models in Vision: A Survey
• Monte Carlo errors with less errors
• Flow-based generative models for Markov chain Monte Carlo in lattice field theory
• Equivariant flow-based sampling for lattice gauge theory
• Topological susceptibility in the SU(3) gauge theory
• Sampling using $SU(N)$ gauge equivariant flows
• Regressive and generative neural networks for scalar field theory
• Diffusion Models in $\text{De Novo}$ Drug Design
• Universality of the topological susceptibility in the SU(3) gauge theory
• Non-Gaussianities in the topological charge distribution of the SU(3) Yang--Mills theory
• Efficient Modelling of Trivializing Maps for Lattice ${\varphi }^4$ Theory Using Normalizing Flows: A First Look at Scalability
• The topological susceptibility in the large-N limit of SU(N) Yang-Mills theory
• Gauge-equivariant flow models for sampling in lattice field theories with pseudofermions
• Topological susceptibility of $N_f=2+1$ QCD from staggered fermions spectral projectors at high temperatures
• Flow-based sampling in the lattice Schwinger model at criticality
• Learning Lattice Quantum Field Theories with Equivariant Continuous Flows
• Reducing Autocorrelation Times in Lattice Simulations with Generative Adversarial Networks
• Advances in machine-learning-based sampling motivated by lattice quantum chromodynamics
• Diffusion Models as Stochastic Quantization in Lattice Field Theory
• Monte Carlo evaluation of the continuum limit of the two-point function of two Euclidean Higgs real scalar fields subject to affine quantization
• Sampling the lattice Nambu-Goto string using Continuous Normalizing Flows
• Numerical determination of the width and shape of the effective string using Stochastic Normalizing Flows
• Scaling of Stochastic Normalizing Flows in $\mathrm{SU}(3)$ lattice gauge theory