Loading…
Automated identification of dominant physical processes
The identification of processes that locally and approximately dominate dynamical system behavior has enabled significant advances in understanding and modeling nonlinear differential dynamical systems. Conventional methods of dominant process identification involve piecemeal and ad hoc (non-rigorou...
Saved in:
| Published in: | Engineering applications of artificial intelligence 2022-11, Vol.116, p.105496, Article 105496 |
|---|---|
| Main Authors: | , , , |
| Format: | Article |
| Language: | English |
| Subjects: | |
| Citations: | Items that this one cites Items that cite this one |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| cited_by | cdi_FETCH-LOGICAL-c387t-6c807d8e6d2c3efd1706a442c66222ce93967d7c16766259ba9c4feee740eeda3 |
|---|---|
| cites | cdi_FETCH-LOGICAL-c387t-6c807d8e6d2c3efd1706a442c66222ce93967d7c16766259ba9c4feee740eeda3 |
| container_end_page | |
| container_issue | |
| container_start_page | 105496 |
| container_title | Engineering applications of artificial intelligence |
| container_volume | 116 |
| creator | Kaiser, Bryan E. Saenz, Juan A. Sonnewald, Maike Livescu, Daniel |
| description | The identification of processes that locally and approximately dominate dynamical system behavior has enabled significant advances in understanding and modeling nonlinear differential dynamical systems. Conventional methods of dominant process identification involve piecemeal and ad hoc (non-rigorous, informal) scaling analyses to identify dominant balances of governing equation terms and to delineate the spatiotemporal boundaries (boundaries in space and/or time) of each dominant balance. For the first time, we present an objective global measure of the fit of dominant balances to observations, which is desirable for automation, and was previously undefined. Furthermore, we propose a formal definition of the dominant balance identification problem in the form of an optimization problem. We show that the optimization can be performed by various machine learning algorithms, enabling the automatic identification of dominant balances. Our method is algorithm agnostic and it eliminates reliance upon expert knowledge to identify dominant balances which are not known beforehand.
[Display omitted]
•Discovery of dominant balances of two or more equation terms yields insights.•Dominant balances often arise in nonlinear dynamical systems.•Automated dominant balance discovery requires hyperparameter tuning.•We present a robust objective criterion for dominant balance identification.•Objective hyper-parameter tuning enables automated dominant balance discovery. |
| doi_str_mv | 10.1016/j.engappai.2022.105496 |
| format | article |
| fullrecord | <record><control><sourceid>elsevier_osti_</sourceid><recordid>TN_cdi_osti_scitechconnect_2290302</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><els_id>S0952197622004869</els_id><sourcerecordid>S0952197622004869</sourcerecordid><originalsourceid>FETCH-LOGICAL-c387t-6c807d8e6d2c3efd1706a442c66222ce93967d7c16766259ba9c4feee740eeda3</originalsourceid><addsrcrecordid>eNqFkE9LAzEQxYMoWKtfQRbvW_Nnm2xulqJVKHjRc4iTWZvSJksShX57d1k9exp4vHlv5kfILaMLRpm83y8wfNq-t37BKeeDuGy0PCMz1ipRSyX1OZlRveQ100pekquc95RS0TZyRtTqq8SjLegq7zAU33mwxcdQxa5y8eiDDaXqd6c86IeqTxEwZ8zX5KKzh4w3v3NO3p8e39bP9fZ187JebWsQrSq1hJYq16J0HAR2jikqbdNwkJJzDqiFlsopYMOZki_1h9XQdIioGororJiTuyk35uJNBl8QdhBDQCiGc00F5YNJTiZIMeeEnemTP9p0MoyakZHZmz9GZmRkJkbD4sO0iMML3x7T2IAB0Pk0Frjo_4v4AYALc3c</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Automated identification of dominant physical processes</title><source>ScienceDirect Freedom Collection</source><creator>Kaiser, Bryan E. ; Saenz, Juan A. ; Sonnewald, Maike ; Livescu, Daniel</creator><creatorcontrib>Kaiser, Bryan E. ; Saenz, Juan A. ; Sonnewald, Maike ; Livescu, Daniel ; Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)</creatorcontrib><description>The identification of processes that locally and approximately dominate dynamical system behavior has enabled significant advances in understanding and modeling nonlinear differential dynamical systems. Conventional methods of dominant process identification involve piecemeal and ad hoc (non-rigorous, informal) scaling analyses to identify dominant balances of governing equation terms and to delineate the spatiotemporal boundaries (boundaries in space and/or time) of each dominant balance. For the first time, we present an objective global measure of the fit of dominant balances to observations, which is desirable for automation, and was previously undefined. Furthermore, we propose a formal definition of the dominant balance identification problem in the form of an optimization problem. We show that the optimization can be performed by various machine learning algorithms, enabling the automatic identification of dominant balances. Our method is algorithm agnostic and it eliminates reliance upon expert knowledge to identify dominant balances which are not known beforehand.
[Display omitted]
•Discovery of dominant balances of two or more equation terms yields insights.•Dominant balances often arise in nonlinear dynamical systems.•Automated dominant balance discovery requires hyperparameter tuning.•We present a robust objective criterion for dominant balance identification.•Objective hyper-parameter tuning enables automated dominant balance discovery.</description><identifier>ISSN: 0952-1976</identifier><identifier>EISSN: 1873-6769</identifier><identifier>DOI: 10.1016/j.engappai.2022.105496</identifier><language>eng</language><publisher>United States: Elsevier Ltd</publisher><subject>Clustering ; Dominant balance ; Dynamical process ; ENGINEERING ; Nonlinear partial differential equations ; scale analysis ; Unsupervised machine learning</subject><ispartof>Engineering applications of artificial intelligence, 2022-11, Vol.116, p.105496, Article 105496</ispartof><rights>2022 Elsevier Ltd</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c387t-6c807d8e6d2c3efd1706a442c66222ce93967d7c16766259ba9c4feee740eeda3</citedby><cites>FETCH-LOGICAL-c387t-6c807d8e6d2c3efd1706a442c66222ce93967d7c16766259ba9c4feee740eeda3</cites><orcidid>0000-0003-2367-1547 ; 0000-0002-4652-6935 ; 0000000262660069 ; 0000000323671547 ; 0000000246526935</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>230,315,786,790,891,27957,27958</link.rule.ids><backlink>$$Uhttps://www.osti.gov/servlets/purl/2290302$$D View this record in Osti.gov$$Hfree_for_read</backlink></links><search><creatorcontrib>Kaiser, Bryan E.</creatorcontrib><creatorcontrib>Saenz, Juan A.</creatorcontrib><creatorcontrib>Sonnewald, Maike</creatorcontrib><creatorcontrib>Livescu, Daniel</creatorcontrib><creatorcontrib>Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)</creatorcontrib><title>Automated identification of dominant physical processes</title><title>Engineering applications of artificial intelligence</title><description>The identification of processes that locally and approximately dominate dynamical system behavior has enabled significant advances in understanding and modeling nonlinear differential dynamical systems. Conventional methods of dominant process identification involve piecemeal and ad hoc (non-rigorous, informal) scaling analyses to identify dominant balances of governing equation terms and to delineate the spatiotemporal boundaries (boundaries in space and/or time) of each dominant balance. For the first time, we present an objective global measure of the fit of dominant balances to observations, which is desirable for automation, and was previously undefined. Furthermore, we propose a formal definition of the dominant balance identification problem in the form of an optimization problem. We show that the optimization can be performed by various machine learning algorithms, enabling the automatic identification of dominant balances. Our method is algorithm agnostic and it eliminates reliance upon expert knowledge to identify dominant balances which are not known beforehand.
[Display omitted]
•Discovery of dominant balances of two or more equation terms yields insights.•Dominant balances often arise in nonlinear dynamical systems.•Automated dominant balance discovery requires hyperparameter tuning.•We present a robust objective criterion for dominant balance identification.•Objective hyper-parameter tuning enables automated dominant balance discovery.</description><subject>Clustering</subject><subject>Dominant balance</subject><subject>Dynamical process</subject><subject>ENGINEERING</subject><subject>Nonlinear partial differential equations</subject><subject>scale analysis</subject><subject>Unsupervised machine learning</subject><issn>0952-1976</issn><issn>1873-6769</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><recordid>eNqFkE9LAzEQxYMoWKtfQRbvW_Nnm2xulqJVKHjRc4iTWZvSJksShX57d1k9exp4vHlv5kfILaMLRpm83y8wfNq-t37BKeeDuGy0PCMz1ipRSyX1OZlRveQ100pekquc95RS0TZyRtTqq8SjLegq7zAU33mwxcdQxa5y8eiDDaXqd6c86IeqTxEwZ8zX5KKzh4w3v3NO3p8e39bP9fZ187JebWsQrSq1hJYq16J0HAR2jikqbdNwkJJzDqiFlsopYMOZki_1h9XQdIioGororJiTuyk35uJNBl8QdhBDQCiGc00F5YNJTiZIMeeEnemTP9p0MoyakZHZmz9GZmRkJkbD4sO0iMML3x7T2IAB0Pk0Frjo_4v4AYALc3c</recordid><startdate>20221101</startdate><enddate>20221101</enddate><creator>Kaiser, Bryan E.</creator><creator>Saenz, Juan A.</creator><creator>Sonnewald, Maike</creator><creator>Livescu, Daniel</creator><general>Elsevier Ltd</general><general>Elsevier</general><scope>AAYXX</scope><scope>CITATION</scope><scope>OIOZB</scope><scope>OTOTI</scope><orcidid>https://orcid.org/0000-0003-2367-1547</orcidid><orcidid>https://orcid.org/0000-0002-4652-6935</orcidid><orcidid>https://orcid.org/0000000262660069</orcidid><orcidid>https://orcid.org/0000000323671547</orcidid><orcidid>https://orcid.org/0000000246526935</orcidid></search><sort><creationdate>20221101</creationdate><title>Automated identification of dominant physical processes</title><author>Kaiser, Bryan E. ; Saenz, Juan A. ; Sonnewald, Maike ; Livescu, Daniel</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c387t-6c807d8e6d2c3efd1706a442c66222ce93967d7c16766259ba9c4feee740eeda3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Clustering</topic><topic>Dominant balance</topic><topic>Dynamical process</topic><topic>ENGINEERING</topic><topic>Nonlinear partial differential equations</topic><topic>scale analysis</topic><topic>Unsupervised machine learning</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Kaiser, Bryan E.</creatorcontrib><creatorcontrib>Saenz, Juan A.</creatorcontrib><creatorcontrib>Sonnewald, Maike</creatorcontrib><creatorcontrib>Livescu, Daniel</creatorcontrib><creatorcontrib>Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)</creatorcontrib><collection>CrossRef</collection><collection>OSTI.GOV - Hybrid</collection><collection>OSTI.GOV</collection><jtitle>Engineering applications of artificial intelligence</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Kaiser, Bryan E.</au><au>Saenz, Juan A.</au><au>Sonnewald, Maike</au><au>Livescu, Daniel</au><aucorp>Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)</aucorp><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Automated identification of dominant physical processes</atitle><jtitle>Engineering applications of artificial intelligence</jtitle><date>2022-11-01</date><risdate>2022</risdate><volume>116</volume><spage>105496</spage><pages>105496-</pages><artnum>105496</artnum><issn>0952-1976</issn><eissn>1873-6769</eissn><notes>LA-UR-22-20119</notes><notes>89233218CNA000001; NA18OAR4320123</notes><notes>USDOE Laboratory Directed Research and Development (LDRD) Program</notes><notes>USDOE National Nuclear Security Administration (NNSA)</notes><notes>National Oceanic and Atmospheric Administration (NOAA)</notes><abstract>The identification of processes that locally and approximately dominate dynamical system behavior has enabled significant advances in understanding and modeling nonlinear differential dynamical systems. Conventional methods of dominant process identification involve piecemeal and ad hoc (non-rigorous, informal) scaling analyses to identify dominant balances of governing equation terms and to delineate the spatiotemporal boundaries (boundaries in space and/or time) of each dominant balance. For the first time, we present an objective global measure of the fit of dominant balances to observations, which is desirable for automation, and was previously undefined. Furthermore, we propose a formal definition of the dominant balance identification problem in the form of an optimization problem. We show that the optimization can be performed by various machine learning algorithms, enabling the automatic identification of dominant balances. Our method is algorithm agnostic and it eliminates reliance upon expert knowledge to identify dominant balances which are not known beforehand.
[Display omitted]
•Discovery of dominant balances of two or more equation terms yields insights.•Dominant balances often arise in nonlinear dynamical systems.•Automated dominant balance discovery requires hyperparameter tuning.•We present a robust objective criterion for dominant balance identification.•Objective hyper-parameter tuning enables automated dominant balance discovery.</abstract><cop>United States</cop><pub>Elsevier Ltd</pub><doi>10.1016/j.engappai.2022.105496</doi><orcidid>https://orcid.org/0000-0003-2367-1547</orcidid><orcidid>https://orcid.org/0000-0002-4652-6935</orcidid><orcidid>https://orcid.org/0000000262660069</orcidid><orcidid>https://orcid.org/0000000323671547</orcidid><orcidid>https://orcid.org/0000000246526935</orcidid><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 0952-1976 |
| ispartof | Engineering applications of artificial intelligence, 2022-11, Vol.116, p.105496, Article 105496 |
| issn | 0952-1976 1873-6769 |
| language | eng |
| recordid | cdi_osti_scitechconnect_2290302 |
| source | ScienceDirect Freedom Collection |
| subjects | Clustering Dominant balance Dynamical process ENGINEERING Nonlinear partial differential equations scale analysis Unsupervised machine learning |
| title | Automated identification of dominant physical processes |
| url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-09-22T04%3A47%3A37IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-elsevier_osti_&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Automated%20identification%20of%20dominant%20physical%20processes&rft.jtitle=Engineering%20applications%20of%20artificial%20intelligence&rft.au=Kaiser,%20Bryan%20E.&rft.aucorp=Los%20Alamos%20National%20Laboratory%20(LANL),%20Los%20Alamos,%20NM%20(United%20States)&rft.date=2022-11-01&rft.volume=116&rft.spage=105496&rft.pages=105496-&rft.artnum=105496&rft.issn=0952-1976&rft.eissn=1873-6769&rft_id=info:doi/10.1016/j.engappai.2022.105496&rft_dat=%3Celsevier_osti_%3ES0952197622004869%3C/elsevier_osti_%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c387t-6c807d8e6d2c3efd1706a442c66222ce93967d7c16766259ba9c4feee740eeda3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |