State predictive information bottleneck
The ability to make sense of the massive amounts of high-dimensional data generated from molecular dynamics simulations is heavily dependent on the knowledge of a low-dimensional manifold (parameterized by a reaction coordinate or RC) that typically distinguishes between relevant metastable states,...
Gespeichert in:
Veröffentlicht in: | The Journal of chemical physics 2021-04, Vol.154 (13), p.134111-134111 |
---|---|
Hauptverfasser: | , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | 134111 |
---|---|
container_issue | 13 |
container_start_page | 134111 |
container_title | The Journal of chemical physics |
container_volume | 154 |
creator | Wang, Dedi Tiwary, Pratyush |
description | The ability to make sense of the massive amounts of high-dimensional data generated from molecular dynamics simulations is heavily dependent on the knowledge of a low-dimensional manifold (parameterized by a reaction coordinate or RC) that typically distinguishes between relevant metastable states, and which captures the relevant slow dynamics of interest. Methods based on machine learning and artificial intelligence have been proposed over the years to deal with learning such low-dimensional manifolds, but they are often criticized for a disconnect from more traditional and physically interpretable approaches. To deal with such concerns, in this work we propose a deep learning based state predictive information bottleneck approach to learn the RC from high-dimensional molecular simulation trajectories. We demonstrate analytically and numerically how the RC learnt in this approach is connected to the committor in chemical physics and can be used to accurately identify transition states. A crucial hyperparameter in this approach is the time delay or how far into the future the algorithm should make predictions about. Through careful comparisons for benchmark systems, we demonstrate that this hyperparameter choice gives useful control over how coarse-grained we want the metastable state classification of the system to be. We thus believe that this work represents a step forward in systematic application of deep learning based ideas to molecular simulations. |
doi_str_mv | 10.1063/5.0038198 |
format | Article |
fullrecord | <record><control><sourceid>proquest_osti_</sourceid><recordid>TN_cdi_osti_scitechconnect_1773945</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2508690644</sourcerecordid><originalsourceid>FETCH-LOGICAL-c511t-d77d03808d90e97e6f851d5704b7f73126779f835ab407856de10c618d9eccba3</originalsourceid><addsrcrecordid>eNp90EtLxDAUBeAgijOOLvwDMujCB3S8aZpHlzL4ggEX6jq0aYod26Ym6YD_3owdFRRcZfNx7slB6BDDDAMjl3QGQAROxRYaYxBpxFkK22gMEOMoZcBGaM-5JQBgHie7aESIIHFM6BidPvrM62lndVEpX630tGpLY5vMV6ad5sb7Wrdave6jnTKrnT7YvBP0fHP9NL-LFg-39_OrRaQoxj4qOC9CFRBFCjrlmpWC4oJySHJecoJjxnlaCkKzPAEuKCs0BsVw8FqpPCMTdDzkGucr6VTltXpRpg0dvMSckzShAZ0NqLPmrdfOy6ZyStd11mrTOxmHLjERYZFAT37RpeltG74QFIiwE0uSoM4HpaxxzupSdrZqMvsuMcj1xJLKzcTBHm0S-7zRxbf82jSAiwGs23_u-G1Wxv4kya4o_8N_T38AMnCOrg</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2508690644</pqid></control><display><type>article</type><title>State predictive information bottleneck</title><source>AIP Journals Complete</source><source>Alma/SFX Local Collection</source><creator>Wang, Dedi ; Tiwary, Pratyush</creator><creatorcontrib>Wang, Dedi ; Tiwary, Pratyush</creatorcontrib><description>The ability to make sense of the massive amounts of high-dimensional data generated from molecular dynamics simulations is heavily dependent on the knowledge of a low-dimensional manifold (parameterized by a reaction coordinate or RC) that typically distinguishes between relevant metastable states, and which captures the relevant slow dynamics of interest. Methods based on machine learning and artificial intelligence have been proposed over the years to deal with learning such low-dimensional manifolds, but they are often criticized for a disconnect from more traditional and physically interpretable approaches. To deal with such concerns, in this work we propose a deep learning based state predictive information bottleneck approach to learn the RC from high-dimensional molecular simulation trajectories. We demonstrate analytically and numerically how the RC learnt in this approach is connected to the committor in chemical physics and can be used to accurately identify transition states. A crucial hyperparameter in this approach is the time delay or how far into the future the algorithm should make predictions about. Through careful comparisons for benchmark systems, we demonstrate that this hyperparameter choice gives useful control over how coarse-grained we want the metastable state classification of the system to be. We thus believe that this work represents a step forward in systematic application of deep learning based ideas to molecular simulations.</description><identifier>ISSN: 0021-9606</identifier><identifier>EISSN: 1089-7690</identifier><identifier>DOI: 10.1063/5.0038198</identifier><identifier>PMID: 33832235</identifier><identifier>CODEN: JCPSA6</identifier><language>eng</language><publisher>United States: American Institute of Physics</publisher><subject>Algorithms ; Artificial intelligence ; Deep learning ; Machine learning ; Metastable state ; Molecular dynamics ; Physics ; Simulation ; Trajectory analysis</subject><ispartof>The Journal of chemical physics, 2021-04, Vol.154 (13), p.134111-134111</ispartof><rights>Author(s)</rights><rights>2021 Author(s). Published under license by AIP Publishing.</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c511t-d77d03808d90e97e6f851d5704b7f73126779f835ab407856de10c618d9eccba3</citedby><cites>FETCH-LOGICAL-c511t-d77d03808d90e97e6f851d5704b7f73126779f835ab407856de10c618d9eccba3</cites><orcidid>0000-0002-2412-6922 ; 0000-0002-7524-5809 ; 0000000224126922 ; 0000000275245809</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://pubs.aip.org/jcp/article-lookup/doi/10.1063/5.0038198$$EHTML$$P50$$Gscitation$$H</linktohtml><link.rule.ids>230,314,776,780,790,881,4497,27903,27904,76130</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/33832235$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink><backlink>$$Uhttps://www.osti.gov/biblio/1773945$$D View this record in Osti.gov$$Hfree_for_read</backlink></links><search><creatorcontrib>Wang, Dedi</creatorcontrib><creatorcontrib>Tiwary, Pratyush</creatorcontrib><title>State predictive information bottleneck</title><title>The Journal of chemical physics</title><addtitle>J Chem Phys</addtitle><description>The ability to make sense of the massive amounts of high-dimensional data generated from molecular dynamics simulations is heavily dependent on the knowledge of a low-dimensional manifold (parameterized by a reaction coordinate or RC) that typically distinguishes between relevant metastable states, and which captures the relevant slow dynamics of interest. Methods based on machine learning and artificial intelligence have been proposed over the years to deal with learning such low-dimensional manifolds, but they are often criticized for a disconnect from more traditional and physically interpretable approaches. To deal with such concerns, in this work we propose a deep learning based state predictive information bottleneck approach to learn the RC from high-dimensional molecular simulation trajectories. We demonstrate analytically and numerically how the RC learnt in this approach is connected to the committor in chemical physics and can be used to accurately identify transition states. A crucial hyperparameter in this approach is the time delay or how far into the future the algorithm should make predictions about. Through careful comparisons for benchmark systems, we demonstrate that this hyperparameter choice gives useful control over how coarse-grained we want the metastable state classification of the system to be. We thus believe that this work represents a step forward in systematic application of deep learning based ideas to molecular simulations.</description><subject>Algorithms</subject><subject>Artificial intelligence</subject><subject>Deep learning</subject><subject>Machine learning</subject><subject>Metastable state</subject><subject>Molecular dynamics</subject><subject>Physics</subject><subject>Simulation</subject><subject>Trajectory analysis</subject><issn>0021-9606</issn><issn>1089-7690</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><recordid>eNp90EtLxDAUBeAgijOOLvwDMujCB3S8aZpHlzL4ggEX6jq0aYod26Ym6YD_3owdFRRcZfNx7slB6BDDDAMjl3QGQAROxRYaYxBpxFkK22gMEOMoZcBGaM-5JQBgHie7aESIIHFM6BidPvrM62lndVEpX630tGpLY5vMV6ad5sb7Wrdave6jnTKrnT7YvBP0fHP9NL-LFg-39_OrRaQoxj4qOC9CFRBFCjrlmpWC4oJySHJecoJjxnlaCkKzPAEuKCs0BsVw8FqpPCMTdDzkGucr6VTltXpRpg0dvMSckzShAZ0NqLPmrdfOy6ZyStd11mrTOxmHLjERYZFAT37RpeltG74QFIiwE0uSoM4HpaxxzupSdrZqMvsuMcj1xJLKzcTBHm0S-7zRxbf82jSAiwGs23_u-G1Wxv4kya4o_8N_T38AMnCOrg</recordid><startdate>20210407</startdate><enddate>20210407</enddate><creator>Wang, Dedi</creator><creator>Tiwary, Pratyush</creator><general>American Institute of Physics</general><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>8FD</scope><scope>H8D</scope><scope>L7M</scope><scope>7X8</scope><scope>OTOTI</scope><orcidid>https://orcid.org/0000-0002-2412-6922</orcidid><orcidid>https://orcid.org/0000-0002-7524-5809</orcidid><orcidid>https://orcid.org/0000000224126922</orcidid><orcidid>https://orcid.org/0000000275245809</orcidid></search><sort><creationdate>20210407</creationdate><title>State predictive information bottleneck</title><author>Wang, Dedi ; Tiwary, Pratyush</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c511t-d77d03808d90e97e6f851d5704b7f73126779f835ab407856de10c618d9eccba3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Algorithms</topic><topic>Artificial intelligence</topic><topic>Deep learning</topic><topic>Machine learning</topic><topic>Metastable state</topic><topic>Molecular dynamics</topic><topic>Physics</topic><topic>Simulation</topic><topic>Trajectory analysis</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Wang, Dedi</creatorcontrib><creatorcontrib>Tiwary, Pratyush</creatorcontrib><collection>PubMed</collection><collection>CrossRef</collection><collection>Technology Research Database</collection><collection>Aerospace Database</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>MEDLINE - Academic</collection><collection>OSTI.GOV</collection><jtitle>The Journal of chemical physics</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Wang, Dedi</au><au>Tiwary, Pratyush</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>State predictive information bottleneck</atitle><jtitle>The Journal of chemical physics</jtitle><addtitle>J Chem Phys</addtitle><date>2021-04-07</date><risdate>2021</risdate><volume>154</volume><issue>13</issue><spage>134111</spage><epage>134111</epage><pages>134111-134111</pages><issn>0021-9606</issn><eissn>1089-7690</eissn><coden>JCPSA6</coden><abstract>The ability to make sense of the massive amounts of high-dimensional data generated from molecular dynamics simulations is heavily dependent on the knowledge of a low-dimensional manifold (parameterized by a reaction coordinate or RC) that typically distinguishes between relevant metastable states, and which captures the relevant slow dynamics of interest. Methods based on machine learning and artificial intelligence have been proposed over the years to deal with learning such low-dimensional manifolds, but they are often criticized for a disconnect from more traditional and physically interpretable approaches. To deal with such concerns, in this work we propose a deep learning based state predictive information bottleneck approach to learn the RC from high-dimensional molecular simulation trajectories. We demonstrate analytically and numerically how the RC learnt in this approach is connected to the committor in chemical physics and can be used to accurately identify transition states. A crucial hyperparameter in this approach is the time delay or how far into the future the algorithm should make predictions about. Through careful comparisons for benchmark systems, we demonstrate that this hyperparameter choice gives useful control over how coarse-grained we want the metastable state classification of the system to be. We thus believe that this work represents a step forward in systematic application of deep learning based ideas to molecular simulations.</abstract><cop>United States</cop><pub>American Institute of Physics</pub><pmid>33832235</pmid><doi>10.1063/5.0038198</doi><tpages>11</tpages><orcidid>https://orcid.org/0000-0002-2412-6922</orcidid><orcidid>https://orcid.org/0000-0002-7524-5809</orcidid><orcidid>https://orcid.org/0000000224126922</orcidid><orcidid>https://orcid.org/0000000275245809</orcidid><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | ISSN: 0021-9606 |
ispartof | The Journal of chemical physics, 2021-04, Vol.154 (13), p.134111-134111 |
issn | 0021-9606 1089-7690 |
language | eng |
recordid | cdi_osti_scitechconnect_1773945 |
source | AIP Journals Complete; Alma/SFX Local Collection |
subjects | Algorithms Artificial intelligence Deep learning Machine learning Metastable state Molecular dynamics Physics Simulation Trajectory analysis |
title | State predictive information bottleneck |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-25T12%3A36%3A52IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_osti_&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=State%20predictive%20information%20bottleneck&rft.jtitle=The%20Journal%20of%20chemical%20physics&rft.au=Wang,%20Dedi&rft.date=2021-04-07&rft.volume=154&rft.issue=13&rft.spage=134111&rft.epage=134111&rft.pages=134111-134111&rft.issn=0021-9606&rft.eissn=1089-7690&rft.coden=JCPSA6&rft_id=info:doi/10.1063/5.0038198&rft_dat=%3Cproquest_osti_%3E2508690644%3C/proquest_osti_%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2508690644&rft_id=info:pmid/33832235&rfr_iscdi=true |