Flow-based generative models for Markov chain Monte Carlo in lattice field theory

Abstract

A Markov chain update scheme using a machine-learned flow-based generative model is proposed for Monte Carlo sampling in lattice field theories. The generative model may be optimized (trained) to produce samples from a distribution approximating the desired Boltzmann distribution determined by the lattice action of the theory being studied. Training the model systematically improves autocorrelation times in the Markov chain, even in regions of parameter space where standard Markov chain Monte Carlo algorithms exhibit critical slowing down in producing decorrelated updates. Moreover, the model may be trained without existing samples from the desired distribution. The algorithm is compared with HMC and local Metropolis sampling for ${\varphi }^4$ theory in two dimensions.

arXiv comment13 pages, 7 figures; corrected normalization conventions in eqns. 20 and 23

References in corpus (29)

• Worm algorithms for classical statistical models
• Monte Carlo errors with less errors
• Parallel WaveNet: Fast High-Fidelity Speech Synthesis
• Self-Learning Monte Carlo Method
• Accelerate Monte Carlo Simulations with Restricted Boltzmann Machines
• Lattice QCD with open boundary conditions and twisted-mass reweighting
• Gauge Equivariant Convolutional Networks and the Icosahedral CNN
• Neural Network Renormalization Group
• Recent Developments of World-Line Monte Carlo Methods
• Regressive and generative neural networks for scalar field theory
• Meron-Cluster Simulation of the $\theta$-Vacuum in the 2-d $O(3)$-Model
• Critical slowing down of topological modes
• Self-Learning Monte Carlo Method in Fermion Systems
• Self-learning Monte Carlo with Deep Neural Networks
• A-NICE-MC: Adversarial Training for MCMC
• Boltzmann Generators -- Sampling Equilibrium States of Many-Body Systems with Deep Learning
• Machine learning action parameters in lattice quantum chromodynamics
• Self-Learning Monte Carlo Method: Continuous-Time Algorithm
• Topological critical slowing down: variations on a toy model
• Monge-Ampère Flow for Generative Modeling
• Exploring the HMC trajectory-length dependence of autocorrelation times in lattice QCD
• Multiscale Monte Carlo equilibration: Pure Yang-Mills theory
• Metadynamics Surfing on Topology Barriers: the $C{P}^{N-1}$ Case
• Lattice QCD on Non-Orientable Manifolds
• Scaling properties of multiscale equilibration
• Testing the event-chain algorithm in asymptotically free models
• Multiscale Monte Carlo equilibration: Two-color QCD with two fermion flavors
• Testing algorithms for critical slowing down
• Thermal Simulations, Open Boundary Conditions and Switches

Cited by in corpus (117)

• Machine learning and the physical sciences
• Anomaly Detection with Density Estimation
• Machine Learning in Nuclear Physics
• Equivariant flow-based sampling for lattice gauge theory
• Machine Learning for Quantum Matter
• AI Poincaré: Machine Learning Conservation Laws from Trajectories
• Estimation of Thermodynamic Observables in Lattice Field Theories with Deep Generative Models
• Sampling using $SU(N)$ gauge equivariant flows
• Asymptotically unbiased estimation of physical observables with neural samplers
• Neural Importance Sampling for Rapid and Reliable Gravitational-Wave Inference
• Exploring QCD matter in extreme conditions with Machine Learning
• FLAG Review 2024
• Precision Studies of QCD in the Low Energy Domain of the EIC
• Targeted free energy estimation via learned mappings
• Nuclear matrix elements from lattice QCD for electroweak and beyond-Standard-Model processes
• Accelerating astronomical and cosmological inference with Preconditioned Monte Carlo
• A Neural Resampler for Monte Carlo Reweighting with Preserved Uncertainties
• Heavy quark potential in quark-gluon Plasma: Deep neural network meets lattice quantum chromodynamics
• A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty
• Efficient Modelling of Trivializing Maps for Lattice ${\varphi }^4$ Theory Using Normalizing Flows: A First Look at Scalability
• Flow-based sampling for fermionic lattice field theories
• Quasi Anomalous Knowledge: Searching for new physics with embedded knowledge
• Phase Transition Study meets Machine Learning
• Neural network reconstruction of the dense matter equation of state from neutron star observables
• Learning Mesh-Based Simulation with Graph Networks
• Gauge-equivariant flow models for sampling in lattice field theories with pseudofermions
• Normalizing flows for atomic solids
• Stochastic normalizing flows as non-equilibrium transformations
• Towards Novel Insights in Lattice Field Theory with Explainable Machine Learning
• Equivariant Flows: sampling configurations for multi-body systems with symmetric energies
• Status and Future Perspectives for Lattice Gauge Theory Calculations to the Exascale and Beyond
• Reliability and Interpretability in Science and Deep Learning
• Unbiased Monte Carlo Cluster Updates with Autoregressive Neural Networks
• Equivariant Hamiltonian Flows
• Ensuring thermodynamic consistency with invertible coarse-graining
• Roadmap on machine learning glassy dynamics
• 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
• Direct sampling of projected entangled-pair states
• Reconstructing spectral functions via automatic differentiation
• Machine-learning physics from unphysics: Finding deconfinement temperature in lattice Yang-Mills theories from outside the scaling window
• Extending machine learning classification capabilities with histogram reweighting
• Learning Trivializing Gradient Flows for Lattice Gauge Theories
• Normalizing Flows and the Real-Time Sign Problem
• Advances in machine-learning-based sampling motivated by lattice quantum chromodynamics
• Ab-initio study of interacting fermions at finite temperature with neural canonical transformation
• Generative models for sampling and phase transition indication in spin systems
• Diffusion Models as Stochastic Quantization in Lattice Field Theory
• Neural Canonical Transformation with Symplectic Flows
• Deep-learning quasi-particle masses from QCD equation of state
• Deep Variational Free Energy Approach to Dense Hydrogen
• Fourier-Flow model generating Feynman paths
• Machine Learning of Thermodynamic Observables in the Presence of Mode Collapse
• Lattice Scalar Field Theory At Complex Coupling
• Free energy calculation of crystalline solids using normalizing flow
• Variational Neural-Network Ansatz for Continuum Quantum Field Theory
• Flow-based density of states for complex actions
• Stochastic Normalizing Flows
• Analysis of autocorrelation times in Neural Markov Chain Monte Carlo simulations
• Sampling the lattice Nambu-Goto string using Continuous Normalizing Flows
• Algorithmic Differentiation for Automated Modeling of Machine Learned Force Fields
• Introduction to Normalizing Flows for Lattice Field Theory
• Differentiable Matrix Elements with MadJax
• Estimating the Euclidean quantum propagator with deep generative modeling of Feynman paths
• Numerical determination of the width and shape of the effective string using Stochastic Normalizing Flows
• Differentiable Physics: A Position Piece
• Hierarchical autoregressive neural networks for statistical systems
• Neural Thermodynamic Integration: Free Energies from Energy-based Diffusion Models
• Efficient mapping of phase diagrams with conditional Boltzmann Generators
• Machine learning a fixed point action for SU(3) gauge theory with a gauge equivariant convolutional neural network
• Generative learning for the problem of critical slowing down in lattice Gross Neveu model
• On learning higher-order cumulants in diffusion models
• Parallel Tempered Metadynamics: Overcoming potential barriers without surfing or tunneling
• Sampling U(1) gauge theory using a re-trainable conditional flow-based model
• Generative Neural Samplers for the Quantum Heisenberg Chain
• Quantum self-learning Monte Carlo with quantum Fourier transform sampler
• Adaptive Monte Carlo augmented with normalizing flows
• Gauge covariant neural network for quarks and gluons
• Generative models for sampling of lattice field theories
• Scaling of Stochastic Normalizing Flows in $\mathrm{SU}(3)$ lattice gauge theory
• Self-Supervised Learning of Generative Spin-Glasses with Normalizing Flows
• Network-Initialized Monte Carlo Based on Generative Neural Networks
• Training normalizing flows with computationally intensive target probability distributions
• Topology changing update algorithms for SU(3) gauge theory
• Simulating first-order phase transition with hierarchical autoregressive networks
• Flow-Based Sampling for Entanglement Entropy and the Machine Learning of Defects
• Temperature-steerable flows
• Self-learning Monte Carlo with equivariant Transformer
• Scaling Up Machine Learning For Quantum Field Theory with Equivariant Continuous Flows
• A Probability Density Theory for Spin-Glass Systems
• Convex optimization and contour deformations
• Nonreversible MCMC from conditional invertible transforms: a complete recipe with convergence guarantees
• Non-Perturbative Trivializing Flows for Lattice Gauge Theories
• Report from the A.I. For Nuclear Physics Workshop
• AdvNF: Reducing Mode Collapse in Conditional Normalising Flows using Adversarial Learning
• Flow-based Nonperturbative Simulation of First-order Phase Transitions
• Scaling flow-based approaches for topology sampling in $\mathrm{SU}(3)$ gauge theory
• Latent Space Refinement for Deep Generative Models
• A Group-Equivariant Autoencoder for Identifying Spontaneously Broken Symmetries
• Sampling in Combinatorial Spaces with SurVAE Flow Augmented MCMC
• Neural Quantum States for Light Nuclei with Chiral Two- and Three-Body Interactions
• Embedded-model flows: Combining the inductive biases of model-free deep learning and explicit probabilistic modeling
• Multi-Lattice Sampling of Quantum Field Theories via Neural Operator-based Flows
• Casimir effect in critical $\mathrm{O}(N)$ models from non-equilibrium Monte Carlo simulations
• Diffusion Models for SU(2) Lattice Gauge Theory in Two Dimensions
• Applications of flow models to the generation of correlated lattice QCD ensembles
• Uncertainty quantification for ptychography using normalizing flows
• Neural Mode Jump Monte Carlo
• A comparative lattice analysis of $SU(2)$ dark gluebal
• Global sampling of Feynman's diagrams through Normalizing Flow
• Hierarchical autoregressive neural networks in three-dimensional statistical system
• NeuMC -- a package for neural sampling for lattice field theories
• Flowing Through Hilbert Space: Quantum-Enhanced Generative Models for Lattice Field Theory
• Implicit Quantile Networks For Emulation in Jet Physics
• Efficient identification of critical regions via Flow Matching-based Monte Carlo initialization
• Group-Equivariant Diffusion Models for Lattice Field Theory