{"id":"https://openalex.org/W7133242594","doi":"https://doi.org/10.52202/068431-0884","title":"In Defense of the Unitary Scalarization for Deep Multi-Task Learning","display_name":"In Defense of the Unitary Scalarization for Deep Multi-Task Learning","publication_year":2022,"publication_date":"2022-01-01","ids":{"openalex":"https://openalex.org/W7133242594","doi":"https://doi.org/10.52202/068431-0884"},"language":"en","primary_location":{"id":"doi:10.52202/068431-0884","doi":"https://doi.org/10.52202/068431-0884","is_oa":false,"landing_page_url":"https://doi.org/10.52202/068431-0884","pdf_url":null,"source":{"id":"https://openalex.org/S4363606243","display_name":"neural information processing systems","issn_l":"1049-5258","issn":["1049-5258"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core","norway-2"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Advances in Neural Information Processing Systems 35","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2201.04122","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5043731054","display_name":"Vitaly Kurin","orcid":"https://orcid.org/0009-0007-9614-8892","observed_orcids":["https://orcid.org/0009-0007-9614-8892"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vitaly Kurin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073957171","display_name":"Alessandro De Palma","orcid":"https://orcid.org/0009-0002-7229-7723","observed_orcids":["https://orcid.org/0009-0002-7229-7723"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Alessandro De Palma","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086657309","display_name":"Ilya Kostrikov","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ilya Kostrikov","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5056879203","display_name":"Shimon Whiteson","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shimon Whiteson","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5078645760","display_name":"Pawan Kumar Mudigonda","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pawan K Mudigonda","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.177,"has_fulltext":false,"cited_by_count":24,"citation_normalized_percentile":{"value":0.94059444,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":93,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"12169","last_page":"12183"},"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":0.8276000022888184,"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":0.8276000022888184,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.07599999755620956,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T12535","display_name":"Machine Learning and Data Classification","score":0.014499999582767487,"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/multi-task-deep-learning","display_name":"multi-task deep learning","score":0.8610000014305115},{"id":"https://openalex.org/keywords/multi-objective-optimization","display_name":"multi-objective optimization","score":0.75},{"id":"https://openalex.org/keywords/scalarization","display_name":"scalarization","score":0.4659999907016754},{"id":"https://openalex.org/keywords/pareto-optimality","display_name":"Pareto optimality","score":0.5120000243186951}],"concepts":[{"id":"https://openalex.org/C67820243","wikidata":"https://www.wikidata.org/wiki/Q179164","display_name":"Unitary state","level":2,"score":0.5738999843597412},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.38659998774528503},{"id":"https://openalex.org/C144237770","wikidata":"https://www.wikidata.org/wiki/Q747534","display_name":"Mathematical economics","level":1,"score":0.37700000405311584},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3546000123023987},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3296999931335449},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.27000001072883606},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.26579999923706055},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.26330000162124634},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2551000118255615},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.2526000142097473},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.52202/068431-0884","doi":"https://doi.org/10.52202/068431-0884","is_oa":false,"landing_page_url":"https://doi.org/10.52202/068431-0884","pdf_url":null,"source":{"id":"https://openalex.org/S4363606243","display_name":"neural information processing systems","issn_l":"1049-5258","issn":["1049-5258"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core","norway-2"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Advances in Neural Information Processing Systems 35","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2201.04122","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2201.04122","pdf_url":"https://arxiv.org/pdf/2201.04122","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":"doi:10.48550/arxiv.2201.04122","doi":"https://doi.org/10.48550/arxiv.2201.04122","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2201.04122","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":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2201.04122","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2201.04122","pdf_url":"https://arxiv.org/pdf/2201.04122","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/G7767666318","display_name":null,"funder_award_id":"EP/L015987/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"}],"funders":[{"id":"https://openalex.org/F4320309480","display_name":"Nvidia","ror":"https://ror.org/03jdj4y14"},{"id":"https://openalex.org/F4320320290","display_name":"University of Oxford","ror":"https://ror.org/052gg0110"},{"id":"https://openalex.org/F4320320300","display_name":"European Commission","ror":"https://ror.org/00k4n6c32"},{"id":"https://openalex.org/F4320332195","display_name":"Samsung","ror":"https://ror.org/04w3jy968"},{"id":"https://openalex.org/F4320334627","display_name":"Engineering and Physical Sciences Research Council","ror":"https://ror.org/0439y7842"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7133252856","https://openalex.org/W7133241239","https://openalex.org/W2890538051","https://openalex.org/W3200842043","https://openalex.org/W4412804409","https://openalex.org/W7196966862","https://openalex.org/W7139926302","https://openalex.org/W4415200242","https://openalex.org/W7133224025","https://openalex.org/W4415014235"],"abstract_inverted_index":{"Recent":[0],"multi-task":[1,20,34,76,94],"learning":[2,83],"research":[3,120],"argues":[4],"against":[5],"unitary":[6,56],"scalarization,":[7,57],"where":[8],"training":[9],"simply":[10],"minimizes":[11],"the":[12,15,72,122],"sum":[13],"of":[14,39,74,102,118],"task":[16],"losses.":[17],"Several":[18],"ad-hoc":[19],"optimization":[21],"algorithms":[22],"have":[23],"instead":[24],"been":[25],"proposed,":[26],"inspired":[27],"by":[28],"various":[29],"hypotheses":[30],"about":[31],"what":[32],"makes":[33],"settings":[35],"difficult.":[36],"The":[37],"majority":[38],"these":[40],"optimizers":[41,77,95],"require":[42],"per-task":[43],"gradients,":[44],"and":[45,50,62,81],"introduce":[46],"significant":[47],"memory,":[48],"runtime,":[49],"implementation":[51],"overhead.":[52],"We":[53,85,109],"show":[54],"that":[55,91],"coupled":[58],"with":[59],"standard":[60],"regularization":[61],"stabilization":[63],"techniques":[64],"from":[65],"single-task":[66],"learning,":[67],"matches":[68],"or":[69],"improves":[70],"upon":[71],"performance":[73],"complex":[75],"in":[78,121],"popular":[79],"supervised":[80],"reinforcement":[82],"settings.":[84],"then":[86],"present":[87],"an":[88],"analysis":[89],"suggesting":[90],"many":[92],"specialized":[93],"can":[96],"be":[97],"partly":[98],"interpreted":[99],"as":[100],"forms":[101],"regularization,":[103],"potentially":[104],"explaining":[105],"our":[106,111],"surprising":[107],"results.":[108],"believe":[110],"results":[112],"call":[113],"for":[114],"a":[115],"critical":[116],"reevaluation":[117],"recent":[119],"area.":[123]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":7},{"year":2024,"cited_by_count":6},{"year":2023,"cited_by_count":8},{"year":2022,"cited_by_count":2}],"updated_date":"2026-10-30T12:39:25.321706","created_date":"2026-03-03T00:00:00"}