{"id":"https://openalex.org/W7161288109","doi":"https://doi.org/10.48550/arxiv.2605.14738","title":"TAPIOCA: Why Task- Aware Pruning Improves OOD model Capability","display_name":"TAPIOCA: Why Task- Aware Pruning Improves OOD model Capability","publication_year":2026,"publication_date":"2026-05-14","ids":{"openalex":"https://openalex.org/W7161288109","doi":"https://doi.org/10.48550/arxiv.2605.14738"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.14738","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.14738","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.14738","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5078456110","display_name":"Krish Sharma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sharma, Krish","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114302110","display_name":"Omar Naim","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Naim, Omar","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5030585541","display_name":"Soumadeep Saha","orcid":"https://orcid.org/0000-0003-0013-5143"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Saha, Soumadeep","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136240510","display_name":"Nicholas Asher","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jain, Vinija","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Chadha, Aman","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chadha, Aman","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Asher, Nicholas","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Asher, Nicholas","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":null,"has_fulltext":false,"cited_by_count":0,"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/T10028","display_name":"Topic Modeling","score":0.23589999973773956,"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/T10028","display_name":"Topic Modeling","score":0.23589999973773956,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.10620000213384628,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.09939999878406525,"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/pruning","display_name":"Pruning","score":0.7695000171661377},{"id":"https://openalex.org/keywords/pairwise-comparison","display_name":"Pairwise comparison","score":0.6212000250816345},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.5654000043869019},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.5041999816894531},{"id":"https://openalex.org/keywords/norm","display_name":"Norm (philosophy)","score":0.42640000581741333},{"id":"https://openalex.org/keywords/regression","display_name":"Regression","score":0.4221000075340271},{"id":"https://openalex.org/keywords/polynomial","display_name":"Polynomial","score":0.3774999976158142},{"id":"https://openalex.org/keywords/data-modeling","display_name":"Data modeling","score":0.3425000011920929}],"concepts":[{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.7695000171661377},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6585000157356262},{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.6212000250816345},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.5654000043869019},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5343000292778015},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.5041999816894531},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4390999972820282},{"id":"https://openalex.org/C191795146","wikidata":"https://www.wikidata.org/wiki/Q3878446","display_name":"Norm (philosophy)","level":2,"score":0.42640000581741333},{"id":"https://openalex.org/C83546350","wikidata":"https://www.wikidata.org/wiki/Q1139051","display_name":"Regression","level":2,"score":0.4221000075340271},{"id":"https://openalex.org/C90119067","wikidata":"https://www.wikidata.org/wiki/Q43260","display_name":"Polynomial","level":2,"score":0.3774999976158142},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3425000011920929},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.33070001006126404},{"id":"https://openalex.org/C28006648","wikidata":"https://www.wikidata.org/wiki/Q6934509","display_name":"Multi-task learning","level":3,"score":0.3100999891757965},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3010999858379364},{"id":"https://openalex.org/C152877465","wikidata":"https://www.wikidata.org/wiki/Q208042","display_name":"Regression analysis","level":2,"score":0.29420000314712524},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.28459998965263367},{"id":"https://openalex.org/C116409475","wikidata":"https://www.wikidata.org/wiki/Q1385056","display_name":"External Data Representation","level":2,"score":0.2840999960899353},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.2797999978065491},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2770000100135803},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.27379998564720154},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.26669999957084656},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2662000060081482},{"id":"https://openalex.org/C120068334","wikidata":"https://www.wikidata.org/wiki/Q45343","display_name":"Polynomial regression","level":3,"score":0.26159998774528503},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.25429999828338623}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.14738","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.14738","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.14738","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.14738","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"work":[1],"has":[2],"promoted":[3],"task-aware":[4,87],"layer":[5],"pruning":[6,46,117],"as":[7,17],"a":[8,83,92,109],"way":[9],"to":[10,82],"improve":[11],"model":[12,172],"performance":[13],"on":[14,50,102,140],"particular":[15],"tasks,":[16],"shown":[18],"by":[19,97,126],"TALE.":[20],"In":[21],"this":[22,124],"paper,":[23],"we":[24],"investigate":[25],"when":[26],"such":[27,45],"improvements":[28],"occur":[29],"and":[30,41,70,134,153,164,167],"why.":[31],"We":[32,60,156],"show":[33,62],"first":[34],"that,":[35],"across":[36,171],"controlled":[37,161],"polynomial":[38],"regression":[39],"tasks":[40],"large":[42],"language":[43],"models,":[44],"yields":[47],"no":[48],"benefit":[49],"in-distribution":[51],"(ID)":[52],"data":[53],"but":[54],"consistently":[55],"improves":[56,154],"out-of-distribution":[57],"(OOD)":[58],"accuracy.":[59],"further":[61],"empirically":[63,96],"that":[64,73,120],"OOD":[65,105,131,146],"inputs":[66,106,147],"induce":[67],"layerwise":[68],"norm":[69],"pairwise-distance":[71],"profiles":[72,100],"deviate":[74],"from":[75],"the":[76,98,113,141,149],"corresponding":[77],"ID":[78,103],"profiles.":[79],"This":[80,144],"leads":[81],"geometric":[84],"explanation":[85],"of":[86,112],"pruning:":[88],"each":[89],"task":[90],"induces":[91],"task-adapted":[93,114,151],"geometry,":[94],"characterized":[95],"representation":[99],"observed":[101,139],"inputs.":[104],"can":[107],"introduce":[108],"distorted":[110],"version":[111],"geometry.":[115],"Task-aware":[116],"identifies":[118],"layers":[119],"create":[121],"or":[122],"amplify":[123],"distortion;":[125],"removing":[127],"them,":[128],"it":[129],"shifts":[130,163],"representational":[132],"norms":[133],"pairwise":[135],"distances":[136],"toward":[137],"those":[138],"adapted":[142],"distribution.":[143],"realigns":[145],"with":[148],"model's":[150],"geometry":[152],"performance.":[155],"provide":[157],"causal":[158],"evidence":[159],"through":[160],"distribution":[162],"residual-scaling":[165],"interventions,":[166],"demonstrate":[168],"consistent":[169],"behavior":[170],"scales.":[173]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-16T00:00:00"}