AI Poincaré: Machine Learning Conservation Laws from Trajectories
Z. Liu, and M. Tegmark. (2020)cite arxiv:2011.04698Comment: 5 pages, 3 figs.
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.
Description
AI Poincar\'e: Machine Learning Conservation Laws from Trajectories
%0 Generic
%1 liu2020poincare
%A Liu, Ziming
%A Tegmark, Max
%D 2020
%K tifr
%T AI Poincaré: Machine Learning Conservation Laws from Trajectories
%U http://arxiv.org/abs/2011.04698
%X 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.
@misc{liu2020poincare,
abstract = {We present AI Poincar\'e, 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.},
added-at = {2020-11-11T07:00:43.000+0100},
author = {Liu, Ziming and Tegmark, Max},
biburl = {https://www.bibsonomy.org/bibtex/22b9cc9b4ada149fbbeadeabaa86420d8/citekhatri},
description = {AI Poincar\'e: Machine Learning Conservation Laws from Trajectories},
interhash = {504fff96b01fca9b0bc382d784fadda8},
intrahash = {2b9cc9b4ada149fbbeadeabaa86420d8},
keywords = {tifr},
note = {cite arxiv:2011.04698Comment: 5 pages, 3 figs},
timestamp = {2020-11-11T07:00:43.000+0100},
title = {AI Poincar\'e: Machine Learning Conservation Laws from Trajectories},
url = {http://arxiv.org/abs/2011.04698},
year = 2020
}