AI Poincaré: Machine Learning Conservation Laws from Trajectories

Abstract

We present AI Poincaré, a machine learning algorithm for auto-discovering conserved quantities using trajectory data from unknown dynamical systems. We test it on five Hamiltonian systems, including the gravitational 3-body problem, and find that it discovers not only all exactly conserved quantities, but also periodic orbits, phase transitions and breakdown timescales for approximate conservation laws.

arXiv commentReplaced by accepted PRL version; expanded validation, improved presentation, more legible figs

References in corpus (19)

• Adam: A Method for Stochastic Optimization
• Searching for Exotic Particles in High-Energy Physics with Deep Learning
• Learning phase transitions by confusion
• Hamiltonian Neural Networks
• Discovering Symbolic Models from Deep Learning with Inductive Biases
• Fast Automated Analysis of Strong Gravitational Lenses with Convolutional Neural Networks
• Flow-based generative models for Markov chain Monte Carlo in lattice field theory
• Integration of Neural Network-Based Symbolic Regression in Deep Learning for Scientific Discovery
• AI Feynman 2.0: Pareto-optimal symbolic regression exploiting graph modularity
• Deep Learning Topological Invariants of Band Insulators
• Lagrangian Neural Networks
• Discovering Symmetry Invariants and Conserved Quantities by Interpreting Siamese Neural Networks
• Physical Symmetries Embedded in Neural Networks
• Extracting Interpretable Physical Parameters from Spatiotemporal Systems using Unsupervised Learning
• Interpretable Conservation Law Estimation by Deriving the Symmetries of Dynamics from Trained Deep Neural Networks
• Learning a Local Symmetry with Neural-Networks
• Learning Symmetries of Classical Integrable Systems
• When effective theories predict: the inevitability of Mercury's anomalous perihelion precession
• Analytical study of chaos and applications

Cited by in corpus (36)

• Deep symbolic regression for physics guided by units constraints: toward the automated discovery of physical laws
• AI Poincaré 2.0: Machine Learning Conservation Laws from Differential Equations
• Translating neutron star observations to nuclear symmetry energy via artificial neural networks
• Machine-Learning Non-Conservative Dynamics for New-Physics Detection
• Discovering Conservation Laws using Optimal Transport and Manifold Learning
• An equation-of-state-meter for CBM using PointNet
• Predicting Beta Decay Energy with Machine Learning
• Foundations of ghost stability
• Machine learning understands knotted polymers
• Modeling assembly bias with machine learning and symbolic regression
• Percent-level constraints on baryonic feedback with spectral distortion measurements
• How close Are Integrable and Non-integrable Models: A Parametric Case Study Based on the Salerno Model
• Kinetic equilibrium of two-dimensional force-free current sheets
• Machine learning prediction for mean motion resonance behaviour -- The planar case
• Metalearning generalizable dynamics from trajectories
• Discover the GellMann-Okubo formula with machine learning
• Learning phase transitions from regression uncertainty: A new regression-based machine learning approach for automated detection of phases of matter
• Discovering Sparse Representations of Lie Groups with Machine Learning
• Extracting Dynamical Frequencies from Invariants of Motion in Finite-Dimensional Nonlinear Integrable Systems
• Analysis of strong coupling constant with machine learning and its application
• Rapidly Encoding Generalizable Dynamics in a Euclidean Symmetric Neural Network
• Machine-assisted discovery of integrable symplectic mappings
• Discovering conservation laws from trajectories via machine learning
• Learning from learning machines: a new generation of AI technology to meet the needs of science
• Regimes of charged particle dynamics in current sheets: the machine learning approach
• Deep learning in bifurcations of particle trajectories
• Using Covariant Lyapunov Vectors to Quantify High Dimensional Chaos with a Conservation Law
• Reveal flocking phase transition of self-propelled active particles by machine learning regression uncertainty
• Intrinsic Dimensionality of Fermi-Pasta-Ulam-Tsingou High-Dimensional Trajectories Through Manifold Learning: A Linear Approach
• Learning the dynamics of Markovian open quantum systems from experimental data
• Topological Learning in Multi-Class Data Sets
• Rediscovery of Numerical Lüscher's Formula from the Neural Network
• Spin-dependent transport in a driven noncolinear antiferromagnetic fractal network
• Deep learning of thermodynamic laws from microscopic dynamics
• Stellar Dynamical Modeling -- Counting Conserved Quantities
• An Information Theory Approach to Physical Domain Discovery