Learning Principle of Least Action with Reinforcement Learning

Nature provides a way to understand physics with reinforcement learning since nature favors the economical way for an object to propagate. In the case of classical mechanics, nature favors the object to move along the path according to the integral of the Lagrangian, called the action $\mathcal{S}$....

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Jin, Zehao, Lin, Joshua Yao-Yu, Li, Siao-Fong
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Jin, Zehao
Lin, Joshua Yao-Yu
Li, Siao-Fong
description Nature provides a way to understand physics with reinforcement learning since nature favors the economical way for an object to propagate. In the case of classical mechanics, nature favors the object to move along the path according to the integral of the Lagrangian, called the action $\mathcal{S}$. We consider setting the reward/penalty as a function of $\mathcal{S}$, so the agent could learn the physical trajectory of particles in various kinds of environments with reinforcement learning. In this work, we verified the idea by using a Q-Learning based algorithm on learning how light propagates in materials with different refraction indices, and show that the agent could recover the minimal-time path equivalent to the solution obtained by Snell's law or Fermat's Principle. We also discuss the similarity of our reinforcement learning approach to the path integral formalism.
doi_str_mv 10.48550/arxiv.2011.11891
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2011_11891</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2011_11891</sourcerecordid><originalsourceid>FETCH-LOGICAL-a671-c3353935f11d09d764c09311bba3d21b7995c162d205c4a22e66ef5abbc8dada3</originalsourceid><addsrcrecordid>eNo1z81qAjEUBeBsXBTtA3RlXmDG3GSSmWwEkaqFgRZxP9z8aUAzEoda377VtqsDB86Bj5AXYGXVSMlmmL_iZ8kZQAnQaHgi89ZjTjHt6UeOycbz0dM-0J_2MtCFHWKf6DUOB7r1MYU-W3_yaaD_qwkZBTxe_PNfjslu9bpbbor2ff22XLQFqhoKK4QUWsgA4Jh2taos0wLAGBSOg6m1lhYUd5xJWyHnXikfJBpjG4cOxZhMf28fgO6c4wnzrbtDugdEfAM0oULp</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Learning Principle of Least Action with Reinforcement Learning</title><source>arXiv.org</source><creator>Jin, Zehao ; Lin, Joshua Yao-Yu ; Li, Siao-Fong</creator><creatorcontrib>Jin, Zehao ; Lin, Joshua Yao-Yu ; Li, Siao-Fong</creatorcontrib><description>Nature provides a way to understand physics with reinforcement learning since nature favors the economical way for an object to propagate. In the case of classical mechanics, nature favors the object to move along the path according to the integral of the Lagrangian, called the action $\mathcal{S}$. We consider setting the reward/penalty as a function of $\mathcal{S}$, so the agent could learn the physical trajectory of particles in various kinds of environments with reinforcement learning. In this work, we verified the idea by using a Q-Learning based algorithm on learning how light propagates in materials with different refraction indices, and show that the agent could recover the minimal-time path equivalent to the solution obtained by Snell's law or Fermat's Principle. We also discuss the similarity of our reinforcement learning approach to the path integral formalism.</description><identifier>DOI: 10.48550/arxiv.2011.11891</identifier><language>eng</language><subject>Computer Science - Learning ; Physics - Classical Physics</subject><creationdate>2020-11</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2011.11891$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2011.11891$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Jin, Zehao</creatorcontrib><creatorcontrib>Lin, Joshua Yao-Yu</creatorcontrib><creatorcontrib>Li, Siao-Fong</creatorcontrib><title>Learning Principle of Least Action with Reinforcement Learning</title><description>Nature provides a way to understand physics with reinforcement learning since nature favors the economical way for an object to propagate. In the case of classical mechanics, nature favors the object to move along the path according to the integral of the Lagrangian, called the action $\mathcal{S}$. We consider setting the reward/penalty as a function of $\mathcal{S}$, so the agent could learn the physical trajectory of particles in various kinds of environments with reinforcement learning. In this work, we verified the idea by using a Q-Learning based algorithm on learning how light propagates in materials with different refraction indices, and show that the agent could recover the minimal-time path equivalent to the solution obtained by Snell's law or Fermat's Principle. We also discuss the similarity of our reinforcement learning approach to the path integral formalism.</description><subject>Computer Science - Learning</subject><subject>Physics - Classical Physics</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNo1z81qAjEUBeBsXBTtA3RlXmDG3GSSmWwEkaqFgRZxP9z8aUAzEoda377VtqsDB86Bj5AXYGXVSMlmmL_iZ8kZQAnQaHgi89ZjTjHt6UeOycbz0dM-0J_2MtCFHWKf6DUOB7r1MYU-W3_yaaD_qwkZBTxe_PNfjslu9bpbbor2ff22XLQFqhoKK4QUWsgA4Jh2taos0wLAGBSOg6m1lhYUd5xJWyHnXikfJBpjG4cOxZhMf28fgO6c4wnzrbtDugdEfAM0oULp</recordid><startdate>20201123</startdate><enddate>20201123</enddate><creator>Jin, Zehao</creator><creator>Lin, Joshua Yao-Yu</creator><creator>Li, Siao-Fong</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20201123</creationdate><title>Learning Principle of Least Action with Reinforcement Learning</title><author>Jin, Zehao ; Lin, Joshua Yao-Yu ; Li, Siao-Fong</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a671-c3353935f11d09d764c09311bba3d21b7995c162d205c4a22e66ef5abbc8dada3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Computer Science - Learning</topic><topic>Physics - Classical Physics</topic><toplevel>online_resources</toplevel><creatorcontrib>Jin, Zehao</creatorcontrib><creatorcontrib>Lin, Joshua Yao-Yu</creatorcontrib><creatorcontrib>Li, Siao-Fong</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Jin, Zehao</au><au>Lin, Joshua Yao-Yu</au><au>Li, Siao-Fong</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Learning Principle of Least Action with Reinforcement Learning</atitle><date>2020-11-23</date><risdate>2020</risdate><abstract>Nature provides a way to understand physics with reinforcement learning since nature favors the economical way for an object to propagate. In the case of classical mechanics, nature favors the object to move along the path according to the integral of the Lagrangian, called the action $\mathcal{S}$. We consider setting the reward/penalty as a function of $\mathcal{S}$, so the agent could learn the physical trajectory of particles in various kinds of environments with reinforcement learning. In this work, we verified the idea by using a Q-Learning based algorithm on learning how light propagates in materials with different refraction indices, and show that the agent could recover the minimal-time path equivalent to the solution obtained by Snell's law or Fermat's Principle. We also discuss the similarity of our reinforcement learning approach to the path integral formalism.</abstract><doi>10.48550/arxiv.2011.11891</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2011.11891
ispartof
issn
language eng
recordid cdi_arxiv_primary_2011_11891
source arXiv.org
subjects Computer Science - Learning
Physics - Classical Physics
title Learning Principle of Least Action with Reinforcement Learning
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-09T18%3A01%3A49IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Learning%20Principle%20of%20Least%20Action%20with%20Reinforcement%20Learning&rft.au=Jin,%20Zehao&rft.date=2020-11-23&rft_id=info:doi/10.48550/arxiv.2011.11891&rft_dat=%3Carxiv_GOX%3E2011_11891%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true