{"id":"https://openalex.org/W7160330567","doi":"https://doi.org/10.48550/arxiv.2605.01959","title":"Flexi-LoRA with Input-Adaptive Ranks: Efficient Finetuning for Speech and Reasoning Tasks","display_name":"Flexi-LoRA with Input-Adaptive Ranks: Efficient Finetuning for Speech and Reasoning Tasks","publication_year":2026,"publication_date":"2026-05-03","ids":{"openalex":"https://openalex.org/W7160330567","doi":"https://doi.org/10.48550/arxiv.2605.01959"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.01959","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.01959","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":null,"license_id":null,"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.2605.01959","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5022488585","display_name":"Zongqian Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Zongqian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080253330","display_name":"Yixuan Su","orcid":"https://orcid.org/0000-0002-1472-7791"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Su, Yixuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135343046","display_name":"Han Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Han","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135312173","display_name":"Zihao Fu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fu, Zihao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135298993","display_name":"Nigel Collier","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Collier, Nigel","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":0.5582000017166138,"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":0.5582000017166138,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.11469999700784683,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.08940000087022781,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/correctness","display_name":"Correctness","score":0.6639000177383423},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5848000049591064},{"id":"https://openalex.org/keywords/consistency","display_name":"Consistency (knowledge bases)","score":0.5658000111579895},{"id":"https://openalex.org/keywords/dependency","display_name":"Dependency (UML)","score":0.5475999712944031},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.5383999943733215},{"id":"https://openalex.org/keywords/rank","display_name":"Rank (graph theory)","score":0.4616999924182892},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.4537999927997589},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4162999987602234}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8240000009536743},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.6639000177383423},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5848000049591064},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.5658000111579895},{"id":"https://openalex.org/C19768560","wikidata":"https://www.wikidata.org/wiki/Q320727","display_name":"Dependency (UML)","level":2,"score":0.5475999712944031},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.5383999943733215},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5037999749183655},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.4616999924182892},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.4537999927997589},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.45210000872612},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4162999987602234},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.3968999981880188},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3767000138759613},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.3465999960899353},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.3310000002384186},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.2815000116825104},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.2703999876976013},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.267300009727478},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2651999890804291},{"id":"https://openalex.org/C12426560","wikidata":"https://www.wikidata.org/wiki/Q189569","display_name":"Basis (linear algebra)","level":2,"score":0.2563999891281128}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.01959","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.01959","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.01959","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.01959","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4","score":0.6540607213973999}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Parameter-efficient":[0],"fine-tuning":[1,189],"methods":[2],"like":[3],"Low-Rank":[4],"Adaptation":[5],"(LoRA)":[6],"have":[7],"become":[8],"essential":[9],"for":[10,22,70,74,182],"deploying":[11],"large":[12],"language":[13],"models,":[14],"yet":[15],"their":[16],"static":[17,134],"parameter":[18,83,165],"allocation":[19,84],"remains":[20],"suboptimal":[21],"inputs":[23],"of":[24,156],"varying":[25],"complexity.":[26,98],"We":[27,171],"present":[28],"Flexi-LoRA,":[29],"a":[30,160,180],"novel":[31],"framework":[32],"that":[33,60,81],"dynamically":[34],"adjusts":[35],"LoRA":[36,135],"ranks":[37],"based":[38],"on":[39,102,145],"input":[40],"complexity":[41],"during":[42],"both":[43],"training":[44,64],"and":[45,55,65,128,187],"inference.":[46],"Through":[47],"empirical":[48,174],"analysis":[49],"across":[50,176],"question":[51,97],"answering,":[52],"mathematical":[53,107],"reasoning,":[54],"speech":[56],"tasks,":[57,178],"we":[58],"demonstrate":[59],"maintaining":[61],"consistency":[62],"between":[63],"inference":[66],"dynamics":[67,104],"is":[68],"important":[69],"effective":[71],"adaptation,":[72],"particularly":[73],"sequential":[75],"reasoning":[76,108,126,149],"tasks.":[77,115],"Our":[78,151],"findings":[79],"reveal":[80],"input-dependent":[82],"achieves":[85],"higher":[86,111],"performance":[87,141],"with":[88,106,140],"fewer":[89,138],"parameters":[90],"by":[91],"optimally":[92],"matching":[93],"rank":[94,103],"configurations":[95],"to":[96],"Furthermore,":[99],"task-specific":[100],"dependency":[101,112],"varies,":[105],"tasks":[109,146],"exhibiting":[110],"than":[113],"QA":[114],"Successful":[116],"adaptation":[117],"manifests":[118],"not":[119],"only":[120],"in":[121,125,185],"correctness":[122],"but":[123],"also":[124],"quality":[127],"instruction":[129],"adherence.":[130],"Flexi-LoRA":[131],"consistently":[132],"outperforms":[133],"while":[136,167],"using":[137],"parameters,":[139],"gains":[142],"more":[143,161],"pronounced":[144],"requiring":[147],"strict":[148],"chains.":[150],"approach":[152],"realizes":[153],"key":[154],"benefits":[155],"mixture-of-experts":[157],"frameworks":[158],"through":[159],"streamlined":[162],"implementation,":[163],"reducing":[164],"redundancy":[166],"improving":[168],"model":[169],"capabilities.":[170],"provide":[172],"comprehensive":[173],"studies":[175],"diverse":[177],"establishing":[179],"basis":[181],"future":[183],"work":[184],"input-adaptive":[186],"efficient":[188],"approaches.":[190]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-06T00:00:00"}