{"id":"https://openalex.org/W7168379722","doi":"https://doi.org/10.48550/arxiv.2607.11940","title":"CARE-LoRA: Compressed Activation REconstruction for Memory-Efficient LoRA","display_name":"CARE-LoRA: Compressed Activation REconstruction for Memory-Efficient LoRA","publication_year":2026,"publication_date":"2026-07-11","ids":{"openalex":"https://openalex.org/W7168379722","doi":"https://doi.org/10.48550/arxiv.2607.11940"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.11940","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.11940","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,"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.2607.11940","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5140760956","display_name":"Gengyu Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Gengyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140912264","display_name":"Haiyin Ran","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ran, Haiyin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140938780","display_name":"Zhengbao He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Zhengbao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140841442","display_name":"Yuhang Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Yuhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140826353","display_name":"Hanling Tian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tian, Hanling","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140785401","display_name":"Zhehao Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Zhehao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5140864524","display_name":"Xiaolin Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Xiaolin","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/T10500","display_name":"Sparse and Compressive Sensing Techniques","score":0.1818999946117401,"subfield":{"id":"https://openalex.org/subfields/2206","display_name":"Computational Mechanics"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10500","display_name":"Sparse and Compressive Sensing Techniques","score":0.1818999946117401,"subfield":{"id":"https://openalex.org/subfields/2206","display_name":"Computational Mechanics"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.17139999568462372,"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":0.15279999375343323,"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/bottleneck","display_name":"Bottleneck","score":0.5375999808311462},{"id":"https://openalex.org/keywords/backpropagation","display_name":"Backpropagation","score":0.5224999785423279},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5162000060081482},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.4643999934196472},{"id":"https://openalex.org/keywords/compressed-sensing","display_name":"Compressed sensing","score":0.46239998936653137},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4381999969482422},{"id":"https://openalex.org/keywords/projection","display_name":"Projection (relational algebra)","score":0.4323999881744385},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.415800005197525}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7720000147819519},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.5375999808311462},{"id":"https://openalex.org/C155032097","wikidata":"https://www.wikidata.org/wiki/Q798503","display_name":"Backpropagation","level":3,"score":0.5224999785423279},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5162000060081482},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.4643999934196472},{"id":"https://openalex.org/C124851039","wikidata":"https://www.wikidata.org/wiki/Q2665459","display_name":"Compressed sensing","level":2,"score":0.46239998936653137},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4381999969482422},{"id":"https://openalex.org/C57493831","wikidata":"https://www.wikidata.org/wiki/Q3134666","display_name":"Projection (relational algebra)","level":2,"score":0.4323999881744385},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.415800005197525},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.41519999504089355},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.4088999927043915},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.3995000123977661},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3736000061035156},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.3714999854564667},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.3357999920845032},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.320499986410141},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.29589998722076416},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.2924000024795532},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.2793000042438507},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.2651999890804291},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.260699987411499},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2526000142097473}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.11940","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.11940","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,"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.2607.11940","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.11940","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,"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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"As":[0],"the":[1,26,46,52,57,65,87,95,100,107,118,133,155],"scale":[2],"of":[3,25,48,91],"large":[4],"pre-trained":[5],"models":[6,146],"continues":[7],"to":[8,131],"grow,":[9],"fine-tuning":[10,31],"them":[11],"under":[12],"limited":[13],"memory":[14,68,157],"budgets":[15],"has":[16],"become":[17],"increasingly":[18],"challenging.":[19],"Low-Rank":[20],"Adaptation":[21],"(LoRA),":[22],"currently":[23],"one":[24],"most":[27],"widely":[28],"adopted":[29],"parameter-efficient":[30],"(PEFT)":[32],"methods,":[33],"mitigates":[34],"this":[35],"challenge":[36],"by":[37,106],"optimizing":[38],"only":[39],"low-rank":[40,101],"adaptation":[41],"matrices,":[42],"thereby":[43,136],"greatly":[44],"reducing":[45,154],"number":[47],"trainable":[49],"parameters.":[50],"With":[51],"parameter":[53],"overhead":[54],"substantially":[55,153],"reduced,":[56],"activations":[58],"retained":[59],"for":[60],"backpropagation":[61,130],"have":[62],"emerged":[63],"as":[64],"primary":[66],"remaining":[67],"bottleneck":[69],"during":[70,117,129],"LoRA":[71,108,138,169,172],"fine-tuning.":[72],"To":[73],"address":[74],"this,":[75],"we":[76],"propose":[77],"CARE-LoRA,":[78],"a":[79,113],"data-aware":[80],"Compressed":[81],"Activation":[82],"REconstruction":[83],"framework.":[84],"By":[85],"exploiting":[86],"inherent":[88],"projection":[89],"structure":[90],"LoRA,":[92],"CARE-LoRA":[93,159],"replaces":[94],"full":[96],"input":[97],"activation":[98,103],"with":[99,121,167],"compressed":[102],"naturally":[104],"produced":[105],"branch.":[109],"It":[110],"further":[111],"computes":[112],"lightweight":[114],"reconstruction":[115],"matrix":[116],"forward":[119],"pass":[120],"negligible":[122],"additional":[123],"computation":[124],"cost,":[125],"which":[126],"is":[127,176],"used":[128],"reconstruct":[132],"gradient":[134],"signal,":[135],"keeping":[137],"matrices":[139],"fully":[140],"trainable.":[141],"Extensive":[142],"experiments":[143],"across":[144],"diverse":[145],"and":[147,170],"downstream":[148],"tasks":[149],"demonstrate":[150],"that,":[151],"while":[152],"overall":[156],"footprint,":[158],"achieves":[160],"competitive":[161],"or":[162],"even":[163],"superior":[164],"performance":[165],"compared":[166],"standard":[168],"representative":[171],"variants.":[173],"Our":[174],"code":[175],"publicly":[177],"available":[178],"at":[179],"https://github.com/fishandyu/CARE-LoRA":[180],".":[181]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-16T00:00:00"}