Roadmap on machine learning glassy dynamics

Abstract

Unraveling the connections between microscopic structure, emergent physical properties, and slow dynamics has long been a challenge when studying the glass transition. The absence of clear visible structural order in amorphous configurations complicates the identification of the key physical mechanisms underpinning slow dynamics. The difficulty in sampling equilibrated configurations at low temperatures hampers thorough numerical and theoretical investigations. This perspective article explores the potential of machine learning (ML) techniques to face these challenges, building on the algorithms that have revolutionized computer vision and image recognition. We present recent successful ML applications, as well as many open problems for the future, such as transferability and interpretability of ML approaches. We highlight new ideas and directions in which ML could provide breakthroughs to better understand the fundamental mechanisms at play in glass-forming liquids. To foster a collaborative community effort, this article also introduces the ``GlassBench" dataset, providing simulation data and benchmarks for both two-dimensional and three-dimensional glass-formers. We propose critical metrics to compare the performance of emerging ML methodologies, in line with benchmarking practices in image and text recognition. The goal of this roadmap is to provide guidelines for the development of ML techniques in systems displaying slow dynamics, while inspiring new directions to improve our theoretical understanding of glassy liquids.

References in corpus (60)

• Theoretical perspective on the glass transition and amorphous materials
• Supercooled Liquids for Pedestrians
• The role of local structure in dynamical arrest
• Irreversible reorganization in a supercooled liquid originates from localised soft modes
• Identifying structural flow defects in disordered solids using machine learning methods
• Discovering Phases, Phase Transitions and Crossovers through Unsupervised Machine Learning: A critical examination
• Combinatorial Optimization with Physics-Inspired Graph Neural Networks
• GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation
• Perspective: Nonequilibrium glassy dynamics in dense systems of active particles
• The Relationship Between Local Structure and Relaxation in Out-of-Equilibrium Glassy Systems
• Structure and dynamics in glass-formers: predictability at large length scales
• Modern computational studies of the glass transition
• Microscopic origin of excess wings in relaxation spectra of supercooled liquids
• Averaging local structure to predict the dynamic propensity in supercooled liquids
• GlassNet: a multitask deep neural network for predicting many glass properties
• Predicting dynamic heterogeneity in glass-forming liquids by physics-inspired machine learning
• Ranking the information content of distance measures
• Elasticity, Facilitation and Dynamic Heterogeneity in Glass-Forming liquids
• An assessment of the structural resolution of various fingerprints commonly used in machine learning
• Elastoplasticity Mediates Dynamical Heterogeneity Below the Mode-Coupling Temperature
• Deep Learning Opacity in Scientific Discovery
• Graph Coloring with Physics-Inspired Graph Neural Networks
• Relevance of Shear Transformations in the Relaxation of Supercooled Liquids
• Tensor-reduced atomic density representations
• BOTAN: BOnd TArgeting Network for prediction of slow glassy dynamics by machine learning relative motion
• Dimensionality reduction of local structure in glassy binary mixtures
• Learning Lattice Quantum Field Theories with Equivariant Continuous Flows
• Skipping the Replica Exchange Ladder with Normalizing Flows
• Finding defects in glasses through machine learning
• Improving the prediction of glassy dynamics by pinpointing the local cage
• Manifolds of quasi-constant SOAP and ACSF fingerprints and the resulting failure to machine learn four-body interactions
• Machine-learning-assisted Monte Carlo fails at sampling computationally hard problems
• Modern graph neural networks do worse than classical greedy algorithms in solving combinatorial optimization problems like maximum independent set
• Sampling with flows, diffusion and autoregressive neural networks: A spin-glass perspective
• Spin glass theory and its new challenge: structured disorder
• Data-driven identification and analysis of the glass transition in polymer melts
• Inability of a graph neural network heuristic to outperform greedy algorithms in solving combinatorial optimization problems like Max-Cut
• Dynamic heterogeneity at the experimental glass transition predicted by transferable machine learning
• Excess wings and asymmetric relaxation spectra in a facilitated trap model
• Discovering dynamic laws from observations: the case of self-propelled, interacting colloids
• Geometry-enhanced graph neural network for learning the smoothness of glassy dynamics from static structure
• Rotation-equivariant Graph Neural Networks for Learning Glassy Liquids Representations
• Exploring glassy dynamics with Markov state models from graph dynamical neural networks
• Normalizing flows as an enhanced sampling method for atomistic supercooled liquids
• Classifying the age of a glass based on structural properties: A machine learning approach
• Policy-guided Monte Carlo on general state spaces: Application to glass-forming mixtures
• Dead or alive: Distinguishing active from passive particles using supervised learning
• Roadmap on Deep Learning for Microscopy
• Selecting Relevant Structural Features for Glassy Dynamics by Information Imbalance
• Building a "trap model" of glassy dynamics from a local structural predictor of rearrangements
• A Boltzmann generator for the isobaric-isothermal ensemble
• Deep reinforced learning heuristic tested on spin-glass ground states: The larger picture
• Rejuvenation and memory effects in active glasses induced by thermal and active cycling
• Unifying framework for strong and fragile liquids via machine learning: a study of liquid silica
• Learning Mappings between Equilibrium States of Liquid Systems Using Normalizing Flows
• Self-Tuning Hamiltonian Monte Carlo for Accelerated Sampling
• Neural networks: from the perceptron to deep nets
• Graph Self-Supervised Learning for Optoelectronic Properties of Organic Semiconductors
• Machine learning-informed structuro-elastoplasticity predicts ductility of disordered solids
• The dynamics of machine-learned "softness" in supercooled liquids describe dynamical heterogeneity

Cited by in corpus (7)

• Graph neural network-based structural classification of glass-forming liquids and its interpretation via self-attention mechanism
• Fluctuations and the limit of predictability in protein evolution
• Active Matter as a framework for living systems-inspired Robophysics
• Computational Methods toward Ultrastable Glasses
• Interpretability of linear regression models of glassy dynamics
• Sampling the Liquid-Gas Critical Point with Boltzmann Generators
• Unsupervised classification of disordered patterns in an oppositely charged colloidal system