{"id":"https://openalex.org/W4387171255","doi":"https://doi.org/10.3233/faia230541","title":"On the Learning Dynamics of Attention Networks","display_name":"On the Learning Dynamics of Attention Networks","publication_year":2023,"publication_date":"2023-09-28","ids":{"openalex":"https://openalex.org/W4387171255","doi":"https://doi.org/10.3233/faia230541"},"language":"en","primary_location":{"id":"doi:10.3233/faia230541","is_oa":true,"landing_page_url":"http://dx.doi.org/10.3233/faia230541","pdf_url":"https://ebooks.iospress.nl/pdf/doi/10.3233/FAIA230541","source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},"type":"book-chapter","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://ebooks.iospress.nl/pdf/doi/10.3233/FAIA230541","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5083753038","display_name":"Rahul Vashisht","orcid":null},"institutions":[{"id":"https://openalex.org/I24676775","display_name":"Indian Institute of Technology Madras","ror":"https://ror.org/03v0r5n49","country_code":"IN","type":"education","lineage":["https://openalex.org/I24676775"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Rahul Vashisht","raw_affiliation_strings":["Department of CSE, Indian Institute of Technology Madras"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of CSE, Indian Institute of Technology Madras","institution_ids":["https://openalex.org/I24676775"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5086266146","display_name":"Harish G. Ramaswamy","orcid":"https://orcid.org/0000-0002-1679-8745"},"institutions":[{"id":"https://openalex.org/I24676775","display_name":"Indian Institute of Technology Madras","ror":"https://ror.org/03v0r5n49","country_code":"IN","type":"education","lineage":["https://openalex.org/I24676775"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Harish G. Ramaswamy","raw_affiliation_strings":["Department of CSE, Indian Institute of Technology Madras"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of CSE, Indian Institute of Technology Madras","institution_ids":["https://openalex.org/I24676775"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I24676775"],"apc_list":null,"apc_paid":null,"fwci":0.4988,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.68015994,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.9855999946594238,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.9855999946594238,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10581","display_name":"Neural dynamics and brain function","score":0.9847999811172485,"subfield":{"id":"https://openalex.org/subfields/2805","display_name":"Cognitive Neuroscience"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10241","display_name":"Functional Brain Connectivity Studies","score":0.9739000201225281,"subfield":{"id":"https://openalex.org/subfields/2805","display_name":"Cognitive Neuroscience"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/initialization","display_name":"Initialization","score":0.7600740790367126},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.7289087176322937},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7123003602027893},{"id":"https://openalex.org/keywords/simple","display_name":"Simple (philosophy)","score":0.5861696004867554},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.5241056680679321},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.51534503698349},{"id":"https://openalex.org/keywords/latent-variable","display_name":"Latent variable","score":0.5013134479522705},{"id":"https://openalex.org/keywords/trajectory","display_name":"Trajectory","score":0.4519979953765869},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.4494398236274719},{"id":"https://openalex.org/keywords/variable","display_name":"Variable (mathematics)","score":0.43853816390037537},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3522633910179138},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.22670739889144897},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.15725788474082947}],"concepts":[{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.7600740790367126},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.7289087176322937},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7123003602027893},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.5861696004867554},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.5241056680679321},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.51534503698349},{"id":"https://openalex.org/C51167844","wikidata":"https://www.wikidata.org/wiki/Q4422623","display_name":"Latent variable","level":2,"score":0.5013134479522705},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.4519979953765869},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4494398236274719},{"id":"https://openalex.org/C182365436","wikidata":"https://www.wikidata.org/wiki/Q50701","display_name":"Variable (mathematics)","level":2,"score":0.43853816390037537},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3522633910179138},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.22670739889144897},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.15725788474082947},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.0},{"id":"https://openalex.org/C1276947","wikidata":"https://www.wikidata.org/wiki/Q333","display_name":"Astronomy","level":1,"score":0.0},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0},{"id":"https://openalex.org/C120665830","wikidata":"https://www.wikidata.org/wiki/Q14620","display_name":"Optics","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.3233/faia230541","is_oa":true,"landing_page_url":"http://dx.doi.org/10.3233/faia230541","pdf_url":"https://ebooks.iospress.nl/pdf/doi/10.3233/FAIA230541","source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"}],"best_oa_location":{"id":"doi:10.3233/faia230541","is_oa":true,"landing_page_url":"http://dx.doi.org/10.3233/faia230541","pdf_url":"https://ebooks.iospress.nl/pdf/doi/10.3233/FAIA230541","source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},"sustainable_development_goals":[{"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10","score":0.5400000214576721}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4387171255.pdf","grobid_xml":"https://content.openalex.org/works/W4387171255.grobid-xml"},"referenced_works_count":26,"referenced_works":["https://openalex.org/W1514535095","https://openalex.org/W1586532344","https://openalex.org/W1810943226","https://openalex.org/W2006969979","https://openalex.org/W2097333193","https://openalex.org/W2133564696","https://openalex.org/W2547062585","https://openalex.org/W2859444450","https://openalex.org/W2889368791","https://openalex.org/W2940906331","https://openalex.org/W2944862027","https://openalex.org/W2950825792","https://openalex.org/W2951025380","https://openalex.org/W2951527505","https://openalex.org/W2952576443","https://openalex.org/W2964190861","https://openalex.org/W2970726176","https://openalex.org/W2977162702","https://openalex.org/W3034636873","https://openalex.org/W3099471123","https://openalex.org/W3115453555","https://openalex.org/W3118608800","https://openalex.org/W3173380736","https://openalex.org/W4288103164","https://openalex.org/W4313447594","https://openalex.org/W4385245566"],"related_works":["https://openalex.org/W3204184292","https://openalex.org/W3176564347","https://openalex.org/W1985458517","https://openalex.org/W2355833770","https://openalex.org/W3031039437","https://openalex.org/W4226299596","https://openalex.org/W4287868543","https://openalex.org/W4309434778","https://openalex.org/W3214759249","https://openalex.org/W4283773090"],"abstract_inverted_index":{"Attention":[0],"models":[1,91],"are":[2,14,32,76],"typically":[3],"learned":[4,92],"by":[5,34],"optimizing":[6],"one":[7],"of":[8,38,50,90,102,105,180,191],"three":[9,30],"standard":[10],"loss":[11,160,183],"functions":[12,184],"that":[13,45,57,176],"variously":[15],"called":[16],"\u2013":[17],"soft":[18,140],"attention,":[19,21],"hard":[20,158],"and":[22,53,82,96,127,150,185,193],"latent":[23],"variable":[24],"marginal":[25],"likelihood":[26],"(LVML)":[27],"attention.":[28],"All":[29],"paradigms":[31,95,122],"motivated":[33],"the":[35,47,51,59,63,71,73,103,106,113,132,139,143,155,163,178,181],"same":[36],"goal":[37],"finding":[39],"two":[40],"models\u2013":[41],"a":[42,54,87,100,124,172,189],"\u2018focus\u2019":[43],"model":[44,56,108,115,145],"\u2018selects\u2019":[46],"right":[48],"segment":[49,61],"input":[52],"\u2018classification\u2019":[55],"processes":[58],"selected":[60,74],"into":[62],"target":[64],"label.":[65],"However,":[66],"they":[67],"differ":[68],"significantly":[69],"in":[70,79,123,162],"way":[72],"segments":[75],"aggregated,":[77],"resulting":[78],"distinct":[80],"dynamics":[81],"final":[83],"results.":[84],"We":[85,118],"observe":[86],"unique":[88],"signature":[89],"using":[93],"these":[94,121],"explain":[97],"this":[98],"as":[99],"consequence":[101],"evolution":[104],"classification":[107],"under":[109,135],"gradient":[110,136],"descent":[111],"when":[112],"focus":[114,144],"is":[116],"fixed.":[117],"also":[119],"analyze":[120],"simple":[125,173],"setting":[126],"derive":[128],"closed-form":[129],"expressions":[130],"for":[131],"parameter":[133],"trajectory":[134],"flow.":[137],"With":[138],"attention":[141,159],"loss,":[142],"improves":[146],"quickly":[147],"at":[148],"initialization":[149],"splutters":[151],"later":[152],"on.":[153],"On":[154],"other":[156],"hand,":[157],"behaves":[161],"opposite":[164],"fashion.":[165],"Based":[166],"on":[167,188],"our":[168],"observations,":[169],"we":[170],"propose":[171],"hybrid":[174],"approach":[175],"combines":[177],"advantages":[179],"different":[182],"demonstrates":[186],"it":[187],"collection":[190],"semi-synthetic":[192],"real-world":[194],"datasets.":[195]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}