Unsupervised machine learning in atomistic simulations, between predictions and understanding

Abstract

Automated analyses of the outcome of a simulation have been an important part of atomistic modeling since the early days, addressing the need of linking the behavior of individual atoms and the collective properties that are usually the final quantity of interest. Methods such as clustering and dimensionality reduction have been used to provide a simplified, coarse-grained representation of the structure and dynamics of complex systems, from proteins to nanoparticles. In recent years, the rise of machine learning has led to an even more widespread use of these algorithms in atomistic modeling, and to consider different classification and inference techniques as part of a coherent toolbox of data-driven approaches. This perspective briefly reviews some of the unsupervised machine-learning methods -- that are geared towards classification and coarse-graining of molecular simulations -- seen in relation to the fundamental mathematical concepts that underlie all machine-learning techniques. It discusses the importance of using concise yet complete representations of atomic structures as the starting point of the analyses, and highlights the risk of introducing preconceived biases when using machine learning to rationalize and understand structure-property relations. Supervised machine-learning techniques, that explicitly attempt to predict the properties of a material given its structure, are less susceptible to such biases. Current developments in the field suggest that using these two classes of approaches side-by-side and in a fully integrated mode, while keeping in mind the relations between the data analysis framework and the fundamental physical principles, will be key for realizing the full potential of machine learning to help understanding the behavior of complex molecules and materials.

References in corpus (36)

• Escaping free-energy minima
• On representing chemical environments
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Two-dimensional materials from high-throughput computational exfoliation of experimentally known compounds
• Ab initio Random Structure Searching
• Identification of slow molecular order parameters for Markov model construction
• SISSO: a compressed-sensing method for identifying the best low-dimensional descriptor in an immensity of offered candidates
• Comparing molecules and solids across structural and alchemical space
• Machine Learning Unifies the Modelling of Materials and Molecules
• AiiDA: Automated Interactive Infrastructure and Database for Computational Science
• From Metadynamics to Dynamics
• How to represent crystal structures for machine learning: towards fast prediction of electronic properties
• Symmetry-Adapted Machine-Learning for Tensorial Properties of Atomistic Systems
• Automatic Selection of Atomic Fingerprints and Reference Configurations for Machine-Learning Potentials
• Materials Cartography: Representing and Mining Material Space Using Structural and Electronic Fingerprints
• Crystal structure prediction using the Minima Hopping method
• Caliber based spectral gap optimization of order parameters (SGOOP) for sampling complex molecular systems
• Understanding molecular representations in machine learning: The role of uniqueness and target similarity
• Accurate molecular polarizabilities with coupled-cluster theory and machine learning
• Atom-Density Representations for Machine Learning
• Accelerated materials property predictions and design using motif-based fingerprints
• Comparison of permutationally invariant polynomials, neural networks, and Gaussian approximation potentials in representing water interactions through many-body expansions
• Efficient nonparametric $n$-body force fields from machine learning
• Metrics for measuring distances in configuration spaces
• Feature Optimization for Atomistic Machine Learning Yields A Data-Driven Construction of the Periodic Table of the Elements
• Predicting polymorphism in molecular crystals using orientational entropy
• iMapD: intrinsic Map Dynamics exploration for uncharted effective free energy landscapes
• A fingerprint based metric for measuring similarities of crystalline structures
• Solid-liquid interface free energy through metadynamics simulations
• Recognizing molecular patterns by machine learning: an agnostic structural definition of the hydrogen bond
• Recognizing Local and Global Structural Motifs at the Atomic Scale
• Free energy surface reconstruction from umbrella samples using Gaussian process regression
• Modeling lithium-ion solid-state electrolytes with a pinball model
• The inefficiency of re-weighted sampling and the curse of system size in high order path integration
• Generalized convex hull construction for materials discovery

Cited by in corpus (36)

• Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
• Big-Data Science in Porous Materials: Materials Genomics and Machine Learning
• Physics-inspired structural representations for molecules and materials
• Perspective on integrating machine learning into computational chemistry and materials science
• Neural Network Potentials for Chemistry: Concepts, Applications and Prospects
• Using Gaussian Process Regression to Simulate the Vibrational Raman Spectra of Molecular Crystals
• SchNetPack 2.0: A neural network toolbox for atomistic machine learning
• Machine Learning for Condensed Matter Physics
• A unified framework for machine learning collective variables for enhanced sampling simulations: $\text{mlcolvar}$
• Efficient implementation of atom-density representations
• Enhanced Sampling in the Age of Machine Learning: Algorithms and Applications
• Driving and characterizing nucleation of urea and glycine polymorphs in water
• Simulating solvation and acidity in complex mixtures with first-principles accuracy: the case of CH${}_3$SO${}_3$H and H${}_2$O${}_2$ in phenol
• A New Kind of Atlas of Zeolite Building Blocks
• Differentiable Molecular Simulations for Control and Learning
• TimeSOAP: Tracking high-dimensional fluctuations in complex molecular systems via time-variations of SOAP spectra
• Ligand additivity relationships enable efficient exploration of transition metal chemical space
• Spectral Map: Embedding Slow Kinetics in Collective Variables
• Manifold Learning in Atomistic Simulations: A Conceptual Review
• Ranking the Synthesizability of Hypothetical Zeolites with the Sorting Hat
• Predicting the Electronic Density Response of Condensed-Phase Systems to Electric Field Perturbations
• Wasserstein metric for improved QML with adjacency matrix representations
• Pattern Matching and Classification of Clusters in Collision Cascades
• Learning the energy curvature versus particle number in approximate density functionals
• Mean-Field Density Matrix Decompositions
• Learning Markovian Dynamics with Spectral Maps
• Molecular Dipole Moment Learning via Rotationally Equivariant Gaussian Process Regression with Derivatives in Molecular-orbital-based Machine Learning
• On the Quality of Uncertainty Estimates from Neural Network Potential Ensembles
• Automated computational workflows for muon spin spectroscopy
• ClasSOMfier: A neural network for cluster analysis and detection of lattice defects
• Explaining classifiers to understand coarse-grained models
• Two-dimensional RMSD projections for reaction path visualization and validation
• A Kernel-based Machine Learning Approach to Computing Quasiparticle Energies within Many-Body Green's Functions Theory
• Global exploration of phase behavior in frustrated Ising models using unsupervised learning techniques
• Ordered ground state configurations of the asymmetric Wigner bilayer system -- revisited: an unsupervised clustering algorithm analysis
• Unsupervised topological learning for identification of atomic structures