{"id":"https://openalex.org/W2995751512","doi":"https://doi.org/10.48550/arxiv.2002.10025","title":"Triple Wins: Boosting Accuracy, Robustness and Efficiency Together by Enabling Input-Adaptive Inference","display_name":"Triple Wins: Boosting Accuracy, Robustness and Efficiency Together by Enabling Input-Adaptive Inference","alternate_titles":[],"publication_year":2020,"publication_date":"2020-02-24","ids":{"openalex":"https://openalex.org/W2995751512","doi":"https://doi.org/10.48550/arxiv.2002.10025","mag":"2995751512"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2002.10025","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2002.10025","pdf_url":"https://arxiv.org/pdf/2002.10025","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"},"parent_works":[],"lineage":["https://openalex.org/W2995751512"],"child_works_count":0,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2002.10025","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5014881080","display_name":"Ting-Kuei Hu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I91045830","display_name":"Texas A&M University","ror":"https://ror.org/01f5ytq51","country_code":"US","type":"education","lineage":["https://openalex.org/I91045830"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Hu, Ting-Kuei","raw_affiliation_strings":["Department of Computer Science and Engineering Texas A&M University , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering Texas A&M University , USA","institution_ids":["https://openalex.org/I91045830"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103073431","display_name":"Tianlong Chen","orcid":"https://orcid.org/0000-0001-7774-8197","observed_orcids":["https://orcid.org/0000-0001-7774-8197"]},"institutions":[{"id":"https://openalex.org/I91045830","display_name":"Texas A&M University","ror":"https://ror.org/01f5ytq51","country_code":"US","type":"education","lineage":["https://openalex.org/I91045830"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Chen, Tianlong","raw_affiliation_strings":["Department of Computer Science and Engineering Texas A&M University , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering Texas A&M University , USA","institution_ids":["https://openalex.org/I91045830"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058253075","display_name":"Haotao Wang","orcid":"https://orcid.org/0000-0002-0626-2058","observed_orcids":["https://orcid.org/0000-0002-0626-2058"]},"institutions":[{"id":"https://openalex.org/I91045830","display_name":"Texas A&M University","ror":"https://ror.org/01f5ytq51","country_code":"US","type":"education","lineage":["https://openalex.org/I91045830"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Wang, Haotao","raw_affiliation_strings":["Department of Computer Science and Engineering Texas A&M University , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering Texas A&M University , USA","institution_ids":["https://openalex.org/I91045830"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5048522863","display_name":"Zhangyang Wang","orcid":"https://orcid.org/0000-0002-2050-5693","observed_orcids":["https://orcid.org/0000-0002-2050-5693"]},"institutions":[{"id":"https://openalex.org/I91045830","display_name":"Texas A&M University","ror":"https://ror.org/01f5ytq51","country_code":"US","type":"education","lineage":["https://openalex.org/I91045830"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Wang, Zhangyang","raw_affiliation_strings":["Department of Computer Science and Engineering Texas A&M University , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering Texas A&M University , USA","institution_ids":["https://openalex.org/I91045830"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I91045830"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":85,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_pulled":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9990000128746033,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9990000128746033,"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":0.0006000000284984708,"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":0.0,"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/adversarial-robustness","display_name":"adversarial robustness","score":0.8859999775886536},{"id":"https://openalex.org/keywords/dynamic-inference","display_name":"dynamic inference","score":0.5239999890327454},{"id":"https://openalex.org/keywords/adversarial-attacks","display_name":"adversarial attacks","score":0.5109999775886536},{"id":"https://openalex.org/keywords/adversarial-defense","display_name":"adversarial defense","score":0.6970000267028809},{"id":"https://openalex.org/keywords/computational-efficiency","display_name":"computational efficiency","score":0.6209999918937683},{"id":"https://openalex.org/keywords/model-capacity","display_name":"model capacity","score":0.46399998664855957}],"concepts":[{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.7829212546348572},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7312777042388916},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7174909710884094},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.5274028182029724},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.5174801349639893},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5026671886444092},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.49289369583129883},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0},{"id":"https://openalex.org/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"pmh:oai:arXiv.org:2002.10025","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2002.10025","pdf_url":"https://arxiv.org/pdf/2002.10025","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":"mag:3008117795","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/pdf/2002.10025.pdf","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},{"id":"doi:10.48550/arxiv.2002.10025","doi":"https://doi.org/10.48550/arxiv.2002.10025","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2002.10025","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:2002.10025","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2002.10025","pdf_url":"https://arxiv.org/pdf/2002.10025","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":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W2995751512.pdf","grobid_xml":"https://content.openalex.org/works/W2995751512.grobid-xml"},"referenced_works_count":46,"referenced_works":["https://openalex.org/W9657784","https://openalex.org/W1845051632","https://openalex.org/W1945616565","https://openalex.org/W2180612164","https://openalex.org/W2194775991","https://openalex.org/W2233116163","https://openalex.org/W2513419314","https://openalex.org/W2516574342","https://openalex.org/W2552767274","https://openalex.org/W2562731582","https://openalex.org/W2604147826","https://openalex.org/W2607219512","https://openalex.org/W2620038827","https://openalex.org/W2640329709","https://openalex.org/W2747261267","https://openalex.org/W2754084392","https://openalex.org/W2755956102","https://openalex.org/W2765233338","https://openalex.org/W2767075075","https://openalex.org/W2767421475","https://openalex.org/W2767962654","https://openalex.org/W2774018344","https://openalex.org/W2788836009","https://openalex.org/W2799032899","https://openalex.org/W2804193803","https://openalex.org/W2808746463","https://openalex.org/W2897355816","https://openalex.org/W2902661385","https://openalex.org/W2907518016","https://openalex.org/W2910426746","https://openalex.org/W2911634294","https://openalex.org/W2913266441","https://openalex.org/W2915594489","https://openalex.org/W2929857780","https://openalex.org/W2951244744","https://openalex.org/W2962677625","https://openalex.org/W2963125010","https://openalex.org/W2963163009","https://openalex.org/W2963382930","https://openalex.org/W2963393494","https://openalex.org/W2963752115","https://openalex.org/W2964153729","https://openalex.org/W2966992648","https://openalex.org/W2971229607","https://openalex.org/W3010833785","https://openalex.org/W3103836116"],"related_works":["https://openalex.org/W4407825874","https://openalex.org/W2887603965","https://openalex.org/W4367598587","https://openalex.org/W3168652588","https://openalex.org/W3126595748","https://openalex.org/W7213423751","https://openalex.org/W3093813382","https://openalex.org/W6967428231","https://openalex.org/W3126379365","https://openalex.org/W4402904010"],"abstract_inverted_index":{"Deep":[0],"networks":[1,111],"were":[2],"recently":[3],"suggested":[4],"to":[5,30,68,95,102,150,165,177,218],"face":[6],"the":[7,34,64,79,155,162,202,219],"odds":[8],"between":[9,73],"accuracy":[10,74,208],"(on":[11,17],"clean":[12,147],"natural":[13],"images)":[14,20],"and":[15,52,75,84,100,131,175,180,209],"robustness":[16,99,130],"adversarially":[18],"perturbed":[19],"(Tsipras":[21],"et":[22,40],"al.,":[23,41],"2019).":[24],"Such":[25],"a":[26,50,60,71,123,186],"dilemma":[27],"is":[28],"shown":[29],"be":[31],"rooted":[32],"in":[33,121,126],"inherently":[35],"higher":[36],"sample":[37],"complexity":[38],"(Schmidt":[39],"2018)":[42],"and/or":[43],"model":[44,65,82,97,128],"capacity":[45,66],"(Nakkiran,":[46],"2019),":[47],"for":[48,89,143],"learning":[49],"high-accuracy":[51],"robust":[53,199],"classifier.":[54],"In":[55],"view":[56],"of":[57,81,154],"that,":[58],"give":[59],"classification":[61],"task,":[62],"growing":[63],"appears":[67],"help":[69],"draw":[70],"win-win":[72],"robustness,":[76,210],"yet":[77,211],"at":[78],"expense":[80],"size":[83],"latency,":[85],"therefore":[86],"posing":[87],"challenges":[88],"resource-constrained":[90],"applications.":[91],"Is":[92],"it":[93],"possible":[94],"co-design":[96],"accuracy,":[98,129],"efficiency":[101],"achieve":[103,206],"their":[104,118],"triple":[105],"wins?":[106],"This":[107],"paper":[108],"studies":[109],"multi-exit":[110],"associated":[112],"with":[113,197,212],"input-adaptive":[114],"efficient":[115],"inference,":[116,201],"showing":[117],"strong":[119],"promise":[120],"achieving":[122],"\"sweet":[124],"point\"":[125],"cooptimizing":[127],"efficiency.":[132],"Our":[133],"proposed":[134],"solution,":[135],"dubbed":[136],"Robust":[137],"Dynamic":[138],"Inference":[139],"Networks":[140],"(RDI-Nets),":[141],"allows":[142],"each":[144],"input":[145],"(either":[146],"or":[148,161],"adversarial)":[149],"adaptively":[151],"choose":[152],"one":[153],"multiple":[156],"output":[157,166],"layers":[158],"(early":[159],"branches":[160],"final":[163],"one)":[164],"its":[167],"prediction.":[168],"That":[169],"multi-loss":[170],"adaptivity":[171],"adds":[172],"new":[173],"variations":[174],"flexibility":[176],"adversarial":[178],"attacks":[179],"defenses,":[181],"on":[182],"which":[183],"we":[184],"present":[185],"systematical":[187],"investigation.":[188],"We":[189],"show":[190],"experimentally":[191],"that":[192],"by":[193],"equipping":[194],"existing":[195],"backbones":[196],"such":[198],"adaptive":[200],"resulting":[203],"RDI-Nets":[204],"can":[205],"better":[207],"over":[213],"30%":[214],"computational":[215],"savings,":[216],"compared":[217],"defended":[220],"original":[221],"models.":[222]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W2995751512","counts_by_year":[{"year":2026,"cited_by_count":6},{"year":2025,"cited_by_count":3},{"year":2024,"cited_by_count":7},{"year":2023,"cited_by_count":17},{"year":2022,"cited_by_count":23},{"year":2021,"cited_by_count":20},{"year":2020,"cited_by_count":9}],"updated_date":"2026-11-20T13:11:24.684918","created_date":"2019-12-26T00:00:00"}
