{"id":"https://openalex.org/W4292829059","doi":"https://doi.org/10.1109/cvprw56347.2022.00295","title":"DA3: Dynamic Additive Attention Adaption for Memory-Efficient On-Device Multi-Domain Learning","display_name":"DA3: Dynamic Additive Attention Adaption for Memory-Efficient On-Device Multi-Domain Learning","publication_year":2022,"publication_date":"2022-06-01","ids":{"openalex":"https://openalex.org/W4292829059","doi":"https://doi.org/10.1109/cvprw56347.2022.00295"},"language":"en","primary_location":{"is_oa":false,"landing_page_url":"https://doi.org/10.1109/cvprw56347.2022.00295","pdf_url":null,"source":{"id":"https://openalex.org/S4363607748","display_name":"2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"version":null,"is_accepted":false,"is_published":false},"type":"article","type_crossref":"proceedings-article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5063655609","display_name":"Yang Li","orcid":"https://orcid.org/0000-0002-2839-6196"},"institutions":[{"id":"https://openalex.org/I55732556","display_name":"Arizona State University","ror":"https://ror.org/03efmqc40","country_code":"US","type":"education","lineage":["https://openalex.org/I3132585960","https://openalex.org/I55732556"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Li Yang","raw_affiliation_string":"Arizona State University,School of Electrical, Computer and Energy Engineering,Tempe,AZ,85287","raw_affiliation_strings":["Arizona State University,School of Electrical, Computer and Energy Engineering,Tempe,AZ,85287"]},{"author_position":"middle","author":{"id":"https://openalex.org/A5030138469","display_name":"Adnan Siraj Rakin","orcid":null},"institutions":[{"id":"https://openalex.org/I55732556","display_name":"Arizona State University","ror":"https://ror.org/03efmqc40","country_code":"US","type":"education","lineage":["https://openalex.org/I3132585960","https://openalex.org/I55732556"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Adnan Siraj Rakin","raw_affiliation_string":"Arizona State University,School of Electrical, Computer and Energy Engineering,Tempe,AZ,85287","raw_affiliation_strings":["Arizona State University,School of Electrical, Computer and Energy Engineering,Tempe,AZ,85287"]},{"author_position":"last","author":{"id":"https://openalex.org/A5047916979","display_name":"Deliang Fan","orcid":"https://orcid.org/0000-0002-7989-6297"},"institutions":[{"id":"https://openalex.org/I55732556","display_name":"Arizona State University","ror":"https://ror.org/03efmqc40","country_code":"US","type":"education","lineage":["https://openalex.org/I3132585960","https://openalex.org/I55732556"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Deliang Fan","raw_affiliation_string":"Arizona State University,School of Electrical, Computer and Energy Engineering,Tempe,AZ,85287","raw_affiliation_strings":["Arizona State University,School of Electrical, Computer and Energy Engineering,Tempe,AZ,85287"]}],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"has_fulltext":false,"cited_by_count":2,"cited_by_percentile_year":{"min":78,"max":83},"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"primary_topic":{"id":"https://openalex.org/T11307","display_name":"Advances in Transfer Learning and Domain Adaptation","score":0.9988,"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":"Advances in Transfer Learning and Domain Adaptation","score":0.9988,"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":"Deep Learning in Computer Vision and Image Recognition","score":0.9985,"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/T12676","display_name":"Theory and Applications of Extreme Learning Machines","score":0.9924,"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":[{"keyword":"dynamic additive attention adaption","score":0.6373},{"keyword":"learning","score":0.3668},{"keyword":"memory-efficient","score":0.25},{"keyword":"on-device","score":0.25},{"keyword":"multi-domain","score":0.25}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7401232},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.73646486},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.6229872},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.5579669},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5560726},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.46935472},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.42469648},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4210168},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.41680452},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3747306},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.20229277},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.12889868},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.091306806},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C120665830","wikidata":"https://www.wikidata.org/wiki/Q14620","display_name":"Optics","level":1,"score":0.0},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"is_oa":false,"landing_page_url":"https://doi.org/10.1109/cvprw56347.2022.00295","pdf_url":null,"source":{"id":"https://openalex.org/S4363607748","display_name":"2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"version":null,"is_accepted":false,"is_published":false}],"best_oa_location":null,"sustainable_development_goals":[],"grants":[{"funder":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation","award_id":null}],"referenced_works_count":11,"referenced_works":["https://openalex.org/W1972420097","https://openalex.org/W2138011018","https://openalex.org/W2194775991","https://openalex.org/W2533598788","https://openalex.org/W2798381792","https://openalex.org/W2804935296","https://openalex.org/W2939059980","https://openalex.org/W2962945654","https://openalex.org/W2963460174","https://openalex.org/W2984612702","https://openalex.org/W3035678286"],"related_works":["https://openalex.org/W4380075502","https://openalex.org/W4223943233","https://openalex.org/W4312200629","https://openalex.org/W4360585206","https://openalex.org/W4364306694","https://openalex.org/W4380086463","https://openalex.org/W4225161397","https://openalex.org/W3014300295","https://openalex.org/W3164822677","https://openalex.org/W4250304930"],"ngrams_url":"https://api.openalex.org/works/W4292829059/ngrams","abstract_inverted_index":{"Nowadays,":[0],"one":[1,22],"practical":[2],"limitation":[3],"of":[4,13,175],"deep":[5],"neural":[6],"network":[7],"(DNN)":[8],"is":[9,49,116],"its":[10],"high":[11,69],"degree":[12],"specialization":[14],"to":[15,28,36,80,213,285],"a":[16,153,165],"single":[17],"task":[18],"or":[19],"domain":[20,138],"(e.g.,":[21,289],"visual":[23],"domain).":[24],"It":[25],"motivates":[26],"re-searchers":[27],"develop":[29],"algorithms":[30],"that":[31,109,119],"can":[32],"adapt":[33],"DNN":[34],"model":[35,179],"multiple":[37,228],"domains":[38],"sequentially,":[39],"while":[40,67,135,171],"still":[41],"performing":[42],"well":[43],"on":[44,60,127,227],"the":[45,117,122,137,173,176,215,251,259,271,286],"past":[46],"domains,":[47],"which":[48,76,233],"known":[50],"as":[51,209],"multi-domain":[52,82,104,157],"learning.":[53],"Almost":[54],"all":[55],"conventional":[56],"methods":[57,288],"only":[58,195],"focus":[59],"improving":[61],"accuracy":[62,140,239],"with":[63],"minimal":[64],"parameter":[65],"update,":[66],"ignoring":[68],"computing":[70],"and":[71,86,125,240,278,293,298],"memory":[72,111,133,198,202,274],"cost":[73,126,217],"during":[74,204],"training,":[75,205],"makes":[77],"it":[78],"difficult":[79],"deploy":[81,245],"learning":[83,158],"into":[84,250],"more":[85,87],"widely":[88],"used":[89,112],"resource-limited":[90],"edge":[91,128,256],"devices,":[92],"like":[93],"mobile":[94],"phone,":[95],"IoT,":[96],"embedded":[97],"system,":[98],"etc.":[99],"During":[100],"our":[101,187,264],"study":[102],"in":[103,237,283],"training":[105,123,132,241,273,279],"process,":[106],"we":[107,142,244],"observe":[108],"large":[110],"for":[113,180,200,218],"activation":[114,197],"storage":[115],"bottleneck":[118],"largely":[120],"limits":[121],"time":[124,280],"devices.":[129],"To":[130],"reduce":[131,214],"usage,":[134],"keeping":[136],"adaption":[139],"performance,":[141],"propose":[143],"Dynamic":[144],"Additive":[145],"Attention":[146],"Adaption":[147],"(DA":[148],"3":[151,163,192,226,249,269],"),":[152],"novel":[154,166],"memory-efficient":[155],"on-device":[156,272],"method.":[159],"DA":[160,189,223,246,266],"learns":[164],"additive":[167],"attention":[168],"adaptor":[169],"module,":[170],"freezing":[172],"weights":[174],"pre-trained":[177],"backbone":[178],"each":[181],"domain.":[182],"Differentiating":[183],"from":[184],"prior":[185],"works,":[186],"proposed":[188,265],"module":[193],"not":[194],"mitigates":[196],"buffering":[199],"reducing":[201],"usage":[203],"but":[206],"also":[207],"serves":[208],"dynamic":[210],"gating":[211],"mechanism":[212],"computation":[216],"fast":[219],"inference.":[220],"We":[221],"validate":[222],"dataset":[229],"against":[230],"state-of-the-art":[231],"methods,":[232],"shows":[234],"great":[235],"improvement":[236],"both":[238],"time.":[242],"Moreover,":[243],"popular":[252],"NIVDIA":[253],"Jetson":[254],"Nano":[255],"GPU,":[257],"where":[258],"measured":[260],"experimental":[261],"results":[262],"show":[263],"reduces":[270],"consumption":[275],"by":[276,281],"5-37\u00d7,":[277],"2\u00d7,":[282],"comparison":[284],"baseline":[287],"standard":[290],"fine-tuning,":[291],"Parallel":[292],"Series":[294],"Res.":[295],"adaptor,":[296],"Piggyback":[297],"TinyTL).":[299]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W4292829059","counts_by_year":[{"year":2023,"cited_by_count":2}],"updated_date":"2024-03-07T18:09:18.702406","created_date":"2022-08-24"}