{"id":"https://openalex.org/W4403780437","doi":"https://doi.org/10.48550/arxiv.2409.14981","title":"On The Specialization of Neural Modules","display_name":"On The Specialization of Neural Modules","publication_year":2024,"publication_date":"2024-09-23","ids":{"openalex":"https://openalex.org/W4403780437","doi":"https://doi.org/10.48550/arxiv.2409.14981"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2409.14981","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2409.14981","pdf_url":"https://arxiv.org/pdf/2409.14981","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2409.14981","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5046091762","display_name":"Devon Jarvis","orcid":"https://orcid.org/0000-0003-2362-7538","observed_orcids":["https://orcid.org/0000-0003-2362-7538"]},"institutions":[{"id":"https://openalex.org/I192619145","display_name":"University of the Witwatersrand","ror":"https://ror.org/03rp50x72","country_code":"ZA","type":"education","lineage":["https://openalex.org/I192619145"]},{"id":"https://openalex.org/I4396570676","display_name":"Sainsbury Wellcome Centre","ror":"https://ror.org/04kjqkz56","country_code":"GB","type":"facility","lineage":["https://openalex.org/I124357947","https://openalex.org/I4396570676","https://openalex.org/I45129253"]},{"id":"https://openalex.org/I4405268824","display_name":"Gatsby Computational Neuroscience Unit","ror":"https://ror.org/0269epf40","country_code":"GB","type":"facility","lineage":["https://openalex.org/I124357947","https://openalex.org/I4405268824","https://openalex.org/I45129253"]},{"id":"https://openalex.org/I45129253","display_name":"University College London","ror":"https://ror.org/02jx3x895","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I45129253"]}],"countries":["GB","ZA"],"is_corresponding":false,"raw_author_name":"Jarvis, Devon","raw_affiliation_strings":["Gatsby Computational Neuroscience Unit & Sainsbury Wellcome Centre , UCL","School of Computer Science and Applied Mathematics , University of the Witwatersrand"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Gatsby Computational Neuroscience Unit & Sainsbury Wellcome Centre , UCL","institution_ids":["https://openalex.org/I4396570676","https://openalex.org/I4405268824","https://openalex.org/I45129253"]},{"raw_affiliation_string":"School of Computer Science and Applied Mathematics , University of the Witwatersrand","institution_ids":["https://openalex.org/I192619145"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102937047","display_name":"Richard Klein","orcid":"https://orcid.org/0000-0003-0783-2072","observed_orcids":["https://orcid.org/0000-0003-0783-2072"]},"institutions":[{"id":"https://openalex.org/I192619145","display_name":"University of the Witwatersrand","ror":"https://ror.org/03rp50x72","country_code":"ZA","type":"education","lineage":["https://openalex.org/I192619145"]}],"countries":["ZA"],"is_corresponding":false,"raw_author_name":"Klein, Richard","raw_affiliation_strings":["School of Computer Science and Applied Mathematics , University of the Witwatersrand"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science and Applied Mathematics , University of the Witwatersrand","institution_ids":["https://openalex.org/I192619145"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068297734","display_name":"Benjamin Rosman","orcid":"https://orcid.org/0000-0002-0284-4114","observed_orcids":["https://orcid.org/0000-0002-0284-4114"]},"institutions":[{"id":"https://openalex.org/I109736498","display_name":"Canadian Institute for Advanced Research","ror":"https://ror.org/01sdtdd95","country_code":"CA","type":"facility","lineage":["https://openalex.org/I109736498"]},{"id":"https://openalex.org/I192619145","display_name":"University of the Witwatersrand","ror":"https://ror.org/03rp50x72","country_code":"ZA","type":"education","lineage":["https://openalex.org/I192619145"]}],"countries":["CA","ZA"],"is_corresponding":false,"raw_author_name":"Rosman, Benjamin","raw_affiliation_strings":["CIFAR Azrieli Global Scholar , CIFAR","School of Computer Science and Applied Mathematics , University of the Witwatersrand"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"CIFAR Azrieli Global Scholar , CIFAR","institution_ids":["https://openalex.org/I109736498"]},{"raw_affiliation_string":"School of Computer Science and Applied Mathematics , University of the Witwatersrand","institution_ids":["https://openalex.org/I192619145"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5011428379","display_name":"Andrew Saxe","orcid":"https://orcid.org/0000-0002-9831-8812","observed_orcids":["https://orcid.org/0000-0002-9831-8812"]},"institutions":[{"id":"https://openalex.org/I109736498","display_name":"Canadian Institute for Advanced Research","ror":"https://ror.org/01sdtdd95","country_code":"CA","type":"facility","lineage":["https://openalex.org/I109736498"]},{"id":"https://openalex.org/I4396570676","display_name":"Sainsbury Wellcome Centre","ror":"https://ror.org/04kjqkz56","country_code":"GB","type":"facility","lineage":["https://openalex.org/I124357947","https://openalex.org/I4396570676","https://openalex.org/I45129253"]},{"id":"https://openalex.org/I4405268824","display_name":"Gatsby Computational Neuroscience Unit","ror":"https://ror.org/0269epf40","country_code":"GB","type":"facility","lineage":["https://openalex.org/I124357947","https://openalex.org/I4405268824","https://openalex.org/I45129253"]},{"id":"https://openalex.org/I45129253","display_name":"University College London","ror":"https://ror.org/02jx3x895","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I45129253"]}],"countries":["CA","GB"],"is_corresponding":false,"raw_author_name":"Saxe, Andrew M.","raw_affiliation_strings":["CIFAR Azrieli Global Scholar , CIFAR","Gatsby Computational Neuroscience Unit & Sainsbury Wellcome Centre , UCL"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"CIFAR Azrieli Global Scholar , CIFAR","institution_ids":["https://openalex.org/I109736498"]},{"raw_affiliation_string":"Gatsby Computational Neuroscience Unit & Sainsbury Wellcome Centre , UCL","institution_ids":["https://openalex.org/I4396570676","https://openalex.org/I4405268824","https://openalex.org/I45129253"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.57669997215271,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.57669997215271,"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/T10320","display_name":"Neural Networks and Applications","score":0.3944999873638153,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.007499999832361937,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/systematic-generalization","display_name":"systematic generalization","score":0.9700000286102295},{"id":"https://openalex.org/keywords/neural-modules","display_name":"neural modules","score":0.8109999895095825},{"id":"https://openalex.org/keywords/compositional-architecture","display_name":"compositional architecture","score":0.8349999785423279},{"id":"https://openalex.org/keywords/systematicity","display_name":"systematicity","score":0.8149999976158142},{"id":"https://openalex.org/keywords/learning-dynamics","display_name":"learning dynamics","score":0.9639999866485596},{"id":"https://openalex.org/keywords/modular-architecture","display_name":"modular architecture","score":0.5009999871253967}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.48699861764907837},{"id":"https://openalex.org/C169760540","wikidata":"https://www.wikidata.org/wiki/Q207011","display_name":"Neuroscience","level":1,"score":0.3692409098148346},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.19123104214668274}],"mesh":[],"locations_count":3,"locations":[{"id":"pmh:oai:arXiv.org:2409.14981","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2409.14981","pdf_url":"https://arxiv.org/pdf/2409.14981","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"pmh:oai:eprints.ucl.ac.uk.OAI2:10169381","doi":null,"is_oa":true,"landing_page_url":"https://discovery.ucl.ac.uk/id/eprint/10169381/","pdf_url":"https://discovery.ucl.ac.uk/id/eprint/10169381/1/5201_on_the_specialization_of_neura.pdf","source":{"id":"https://openalex.org/S4306400024","display_name":"UCL Discovery (University College London)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I45129253","host_organization_name":"University College London","host_organization_lineage":["https://openalex.org/I45129253"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"In: Proceedings of the Eleventh International Conference on Learning Representations. (pp. pp. 1-31). ICLR (2023) (In press).","raw_type":"Proceedings paper"},{"id":"doi:10.48550/arxiv.2409.14981","doi":"https://doi.org/10.48550/arxiv.2409.14981","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2409.14981","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2409.14981","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2409.14981","pdf_url":"https://arxiv.org/pdf/2409.14981","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G1345892341","display_name":"Sainsbury Wellcome Centre Core Grant Renewal","funder_award_id":"219627/Z/19/Z","funder_id":"https://openalex.org/F4320311904","funder_display_name":"Wellcome Trust"},{"id":"https://openalex.org/G9000106608","display_name":null,"funder_award_id":"216386/Z/19/Z","funder_id":"https://openalex.org/F4320311904","funder_display_name":"Wellcome Trust"}],"funders":[{"id":"https://openalex.org/F4320309949","display_name":"Canadian Institute for Advanced Research","ror":"https://ror.org/01sdtdd95"},{"id":"https://openalex.org/F4320311904","display_name":"Wellcome Trust","ror":"https://ror.org/029chgv08"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4403780437.pdf","grobid_xml":"https://content.openalex.org/works/W4403780437.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"A":[0],"number":[1],"of":[2,12,26,60,78,101,112,119,126,133,143,157],"machine":[3],"learning":[4,124],"models":[5,30],"have":[6],"been":[7],"proposed":[8],"with":[9,54],"the":[10,16,58,67,76,123,134,141,155,167],"goal":[11],"achieving":[13],"systematic":[14,91,106],"generalization:":[15],"ability":[17,77],"to":[18,36,41,50,81,83,151,160,175],"reason":[19],"about":[20],"new":[21],"situations":[22],"by":[23,65,104],"combining":[24],"aspects":[25],"previous":[27],"experiences.":[28],"These":[29],"leverage":[31],"compositional":[32],"architectures":[33,62,159],"which":[34],"aim":[35],"learn":[37],"specialized":[38],"modules":[39,68,80,129,150],"dedicated":[40],"structures":[42,85],"in":[43,86,170],"a":[44,87,98,116],"task":[45],"that":[46,166],"can":[47],"be":[48],"composed":[49],"solve":[51],"novel":[52],"problems":[53],"similar":[55],"structures.":[56],"While":[57],"compositionality":[59],"these":[61],"is":[63,70,147],"guaranteed":[64],"design,":[66],"specializing":[69],"not.":[71],"Here":[72],"we":[73,96,114,164],"theoretically":[74],"study":[75,122],"network":[79],"specialize":[82],"useful":[84],"dataset":[88],"and":[89,121,154,179],"achieve":[90,161],"generalization.":[92],"To":[93],"this":[94,110],"end":[95],"introduce":[97],"minimal":[99],"space":[100,111],"datasets":[102,113,178],"motivated":[103],"practical":[105],"generalization":[107],"benchmarks.":[108],"From":[109],"present":[115],"mathematical":[117],"definition":[118],"systematicity":[120],"dynamics":[125],"linear":[127],"neural":[128],"when":[130],"solving":[131],"components":[132],"task.":[135],"Our":[136],"results":[137,169],"shed":[138],"light":[139],"on":[140],"difficulty":[142],"module":[144],"specialization,":[145],"what":[146],"required":[148],"for":[149],"successfully":[152],"specialize,":[153],"necessity":[156],"modular":[158],"systematicity.":[162],"Finally,":[163],"confirm":[165],"theoretical":[168],"our":[171],"tractable":[172],"setting":[173],"generalize":[174],"more":[176],"complex":[177],"non-linear":[180],"architectures.":[181]},"counts_by_year":[{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":1}],"updated_date":"2026-10-16T09:59:22.235173","created_date":"2025-10-10T00:00:00"}