Normalizing flows as an enhanced sampling method for atomistic supercooled liquids

Abstract

Normalizing flows can transform a simple prior probability distribution into a more complex target distribution. Here, we evaluate the ability and efficiency of generative machine learning methods to sample the Boltzmann distribution of an atomistic model for glass-forming liquids. This is a notoriously difficult task, as it amounts to ergodically exploring the complex free energy landscape of a disordered and frustrated many-body system. We optimize a normalizing flow model to successfully transform high-temperature configurations of a dense liquid into low-temperature ones, near the glass transition. We perform a detailed comparative analysis with established enhanced sampling techniques developed in the physics literature to assess and rank the performance of normalizing flows against state-of-the-art algorithms. We demonstrate that machine learning methods are very promising, showing a large speedup over conventional molecular dynamics. Normalizing flows show performances comparable to parallel tempering and population annealing, while still falling far behind the swap Monte Carlo algorithm. Our study highlights the potential of generative machine learning models in scientific computing for complex systems, but also points to some of its current limitations and the need for further improvement.

References in corpus (58)

• Auto-Encoding Variational Bayes
• Denoising Diffusion Probabilistic Models
• Generative Adversarial Networks
• Exchange Monte Carlo Method and Application to Spin Glass Simulations
• Theoretical perspective on the glass transition and amorphous materials
• Simulated Tempering: A New Monte Carlo Scheme
• Deep Unsupervised Learning using Nonequilibrium Thermodynamics
• Variational Inference with Normalizing Flows
• Testing Mode-Coupling Theory for a Supercooled Binary Lennard-Jones Mixture I: The van Hove Correlation Function
• Parallel Tempering: Theory, Applications, and New Perspectives
• Models and algorithms for the next generation of glass transition studies
• A Biased Monte Carlo Scheme for Zeolite Structure Solution
• Solving Statistical Mechanics Using Variational Autoregressive Networks
• A Physics-informed Diffusion Model for High-fidelity Flow Field Reconstruction
• Event-chain Monte Carlo algorithms for hard-sphere systems
• Fast Monte Carlo algorithm for supercooled soft spheres
• GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation
• Equilibrium sampling of hard spheres up to the jamming density and beyond
• Probing a liquid to glass transition in equilibrium
• Thirty milliseconds in the life of a supercooled liquid
• Flow Matching for Generative Modeling
• Modern computational studies of the glass transition
• Population Annealing with Weighted Averages: A Monte Carlo Method for Rough Free Energy Landscapes
• Replica-exchange molecular dynamics simulation for supercooled liquids
• Parallel Excluded Volume Tempering for Polymer Melts
• Efficient swap algorithms for molecular dynamics simulations of equilibrium supercooled liquids
• A-NICE-MC: Adversarial Training for MCMC
• Assessing the structural heterogeneity of supercooled liquids through community inference
• Population annealing: Theory and application in spin glasses
• Equivariant Flows: Exact Likelihood Generative Learning for Symmetric Densities
• Predicting dynamic heterogeneity in glass-forming liquids by physics-inspired machine learning
• Boosting Monte Carlo simulations of spin glasses using autoregressive neural networks
• Use the force! Reduced variance estimators for densities, radial distribution functions and local mobilities in molecular simulations
• Point-to-set lengths, local structure, and glassiness
• Ultrastable metallic glasses in silico
• Stochastic Normalizing Flows
• Efficient measurement of point-to-set correlations and overlap fluctuations in glass-forming liquids
• Skipping the Replica Exchange Ladder with Normalizing Flows
• Towards Predicting Equilibrium Distributions for Molecular Systems with Deep Learning
• Machine-learning-assisted Monte Carlo fails at sampling computationally hard problems
• Sampling efficiency of transverse forces in dense liquids
• Efficient Monte Carlo simulation of a glass forming binary mixture
• Sampling with flows, diffusion and autoregressive neural networks: A spin-glass perspective
• Analysis and Optimization of Population Annealing
• Irreversible Monte Carlo algorithms for hard disk glasses: from event-chain to collective swaps
• Equilibration times in numerical simulation of structural glasses: Comparing parallel tempering and conventional molecular dynamics
• Timewarp: Transferable Acceleration of Molecular Dynamics by Learning Time-Coarsened Dynamics
• Dynamic heterogeneity at the experimental glass transition predicted by transferable machine learning
• Computer simulations of the glass transition and glassy materials
• Flow Annealed Importance Sampling Bootstrap
• Accelerating equilibrium spin-glass simulations using quantum annealers via generative deep learning
• Single-parameter aging in a binary Lennard-Jones system
• Equivariant flow matching
• A Boltzmann generator for the isobaric-isothermal ensemble
• Learning Mappings between Equilibrium States of Liquid Systems Using Normalizing Flows
• SE(3) Equivariant Augmented Coupling Flows
• Designing losses for data-free training of normalizing flows on Boltzmann distributions
• Measuring glass entropies with population annealing

Cited by in corpus (9)

• Roadmap on machine learning glassy dynamics
• Monte Carlo simulations of glass-forming liquids beyond Metropolis
• Policy-guided Monte Carlo on general state spaces: Application to glass-forming mixtures
• Graph neural network-based structural classification of glass-forming liquids and its interpretation via self-attention mechanism
• Characterising the slow dynamics of the swap Monte Carlo algorithm
• Irreversible swap algorithms for soft sphere glasses
• Molecular motion at the experimental glass transition
• Computational Methods toward Ultrastable Glasses
• Sampling the Liquid-Gas Critical Point with Boltzmann Generators