Neural Symplectic Integrator with Hamiltonian Inductive Bias for the Gravitational $N$-body Problem

Abstract

The gravitational $N$-body problem, which is fundamentally important in astrophysics to predict the motion of $N$ celestial bodies under the mutual gravity of each other, is usually solved numerically because there is no known general analytical solution for $N>2$. Can an $N$-body problem be solved accurately by a neural network (NN)? Can a NN observe long-term conservation of energy and orbital angular momentum? Inspired by Wistom & Holman (1991)'s symplectic map, we present a neural $N$-body integrator for splitting the Hamiltonian into a two-body part, solvable analytically, and an interaction part that we approximate with a NN. Our neural symplectic $N$-body code integrates a general three-body system for $10^5$ steps without diverting from the ground truth dynamics obtained from a traditional $N$-body integrator. Moreover, it exhibits good inductive bias by successfully predicting the evolution of $N$-body systems that are no part of the training set.

arXiv comment7 pages, 2 figures, accepted for publication at the NeurIPS 2021 workshop "Machine Learning and the Physical Sciences"

References in corpus (15)

• Adam: A Method for Stochastic Optimization
• Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations
• The Eccentric Kozai-Lidov Effect and Its Applications
• Temperate Earth-sized planets transiting a nearby ultracool dwarf star
• Scalable Training of Artificial Neural Networks with Adaptive Sparse Connectivity inspired by Network Science
• Hamiltonian Neural Networks
• Symplectic ODE-Net: Learning Hamiltonian Dynamics with Control
• Hamiltonian Graph Networks with ODE Integrators
• Symplectic Recurrent Neural Networks
• Lagrangian Neural Networks
• Detecting outliers and learning complex structures with large spectroscopic surveys - a case study with APOGEE stars
• HIGAN: Cosmic Neutral Hydrogen with Generative Adversarial Networks
• Hamiltonian Generative Networks
• SNIascore: Deep Learning Classification of Low-Resolution Supernova Spectra
• Simplifying Hamiltonian and Lagrangian Neural Networks via Explicit Constraints