Deep learning of thermodynamic laws from microscopic dynamics

Abstract

We numerically show that a deep neural network (DNN) can learn macroscopic thermodynamic laws purely from microscopic data. Using molecular dynamics simulations, we generate the data of snapshot images of gas particles undergoing adiabatic processes. We train a DNN to determine the temporal order of input image pairs. We observe that the trained network induces an order relation between states consistent with adiabatic accessibility, satisfying the axioms of thermodynamics. Furthermore, the internal representation learned by the DNN act as an entropy. These results suggest that machine learning can discover emergent physical laws that are valid at scales far larger than those of the underlying constituents -- opening a pathway to data-driven discovery of macroscopic physics.

10 pages, 9 figures

References in corpus (36)

• Machine learning phases of matter
• Colloquium. Quantum Fluctuation Relations: Foundations and Applications
• Discovering Phase Transitions with Unsupervised Learning
• The Physics and Mathematics of the Second Law of Thermodynamics
• Unsupervised learning of phase transitions: from principal component analysis to variational autoencoders
• Discovering physical concepts with neural networks
• Machine learning quantum phases of matter beyond the fermion sign problem
• Interpretable Scientific Discovery with Symbolic Regression: A Review
• Hard-disk equation of state: First-order liquid-hexatic transition in two dimensions with three simulation methods
• Learning Thermodynamics with Boltzmann Machines
• Phase diagram and universality of the Lennard-Jones gas-liquid system
• Detection of phase transition via convolutional neural network
• Unsupervised machine learning account of magnetic transitions in the Hubbard model
• Machine Learning of Explicit Order Parameters: From the Ising Model to SU(2) Lattice Gauge Theory
• AI Poincaré: Machine Learning Conservation Laws from Trajectories
• Deep Learning the Quantum Phase Transitions in Random Two-Dimensional Electron Systems
• Machine learning of phase transitions in the percolation and XY models
• A computational framework for physics-informed symbolic regression with straightforward integration of domain knowledge
• Machine learning of quantum phase transitions
• Unveiling phase transitions with machine learning
• Heat conduction in a three dimensional anharmonic crystal
• Physics enhanced neural networks predict order and chaos
• Scale-invariant Feature Extraction of Neural Network and Renormalization Group Flow
• Discovering Symmetry Invariants and Conserved Quantities by Interpreting Siamese Neural Networks
• Machine-learning hidden symmetries
• The entropy concept for non-equilibrium states
• Symbolic Pregression: Discovering Physical Laws from Distorted Video
• Thermodynamics and Feature Extraction by Machine Learning
• Adaptable Hamiltonian neural networks
• Phase transition encoded in neural network
• Interpretable Conservation Law Estimation by Deriving the Symmetries of Dynamics from Trained Deep Neural Networks
• Machine-learning detection of the Berezinskii-Kosterlitz-Thouless transition and the second-order phase transition in the XXZ models
• Machine learning of phases and structures for model systems in physics
• Logical Reasoning for Revealing the Critical Temperature through Deep Learning of Configuration Ensemble of Statistical Systems
• A Guide to Entropy and the Second Law of Thermodynamics
• Machine-Learning Detection of the Berezinskii-Kosterlitz-Thouless Transitions