{"id":"https://openalex.org/W3175192640","doi":"https://doi.org/10.1109/cvpr46437.2021.01607","title":"See through Gradients: Image Batch Recovery via GradInversion","display_name":"See through Gradients: Image Batch Recovery via GradInversion","publication_year":2021,"publication_date":"2021-06-01","ids":{"openalex":"https://openalex.org/W3175192640","doi":"https://doi.org/10.1109/cvpr46437.2021.01607","mag":"3175192640"},"language":"en","primary_location":{"id":"doi:10.1109/cvpr46437.2021.01607","is_oa":false,"landing_page_url":"https://doi.org/10.1109/cvpr46437.2021.01607","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","raw_type":"proceedings-article"},"type":"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/A5002444694","display_name":"Hongxu Yin","orcid":"https://orcid.org/0000-0002-6481-6389"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Hongxu Yin","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017650914","display_name":"Arun Mallya","orcid":"https://orcid.org/0000-0002-1348-4740"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Arun Mallya","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038984764","display_name":"Arash Vahdat","orcid":"https://orcid.org/0009-0005-9476-1306"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Arash Vahdat","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101540588","display_name":"Jos\u00e9 M. Alvarez","orcid":"https://orcid.org/0000-0002-7535-6322"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jose M. Alvarez","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5056503617","display_name":"Jan Kautz","orcid":"https://orcid.org/0000-0002-8830-429X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jan Kautz","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5066945976","display_name":"Pavlo Molchanov","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pavlo Molchanov","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5002444694"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":41.2955,"has_fulltext":false,"cited_by_count":413,"citation_normalized_percentile":{"value":0.99866442,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":89,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"16332","last_page":"16341"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9994000196456909,"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":0.9994000196456909,"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":0.9993000030517578,"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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.9987999796867371,"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/computer-science","display_name":"Computer science","score":0.7600241899490356},{"id":"https://openalex.org/keywords/regularization","display_name":"Regularization (linguistics)","score":0.6761027574539185},{"id":"https://openalex.org/keywords/fidelity","display_name":"Fidelity","score":0.5710185170173645},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.5670874118804932},{"id":"https://openalex.org/keywords/consistency","display_name":"Consistency (knowledge bases)","score":0.5196458101272583},{"id":"https://openalex.org/keywords/encode","display_name":"ENCODE","score":0.4841485619544983},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.4841104745864868},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.4795079529285431},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.4701792299747467},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.46639224886894226},{"id":"https://openalex.org/keywords/minification","display_name":"Minification","score":0.4608878791332245},{"id":"https://openalex.org/keywords/point","display_name":"Point (geometry)","score":0.43082401156425476},{"id":"https://openalex.org/keywords/class","display_name":"Class (philosophy)","score":0.42684298753738403},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.41808751225471497},{"id":"https://openalex.org/keywords/data-mining","display_name":"Data mining","score":0.34078314900398254},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.1645711064338684},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.09586408734321594}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7600241899490356},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.6761027574539185},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.5710185170173645},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.5670874118804932},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.5196458101272583},{"id":"https://openalex.org/C66746571","wikidata":"https://www.wikidata.org/wiki/Q1134833","display_name":"ENCODE","level":3,"score":0.4841485619544983},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.4841104745864868},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4795079529285431},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.4701792299747467},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.46639224886894226},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.4608878791332245},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.43082401156425476},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.42684298753738403},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.41808751225471497},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.34078314900398254},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.1645711064338684},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.09586408734321594},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"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/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/cvpr46437.2021.01607","is_oa":false,"landing_page_url":"https://doi.org/10.1109/cvpr46437.2021.01607","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Partnerships for the goals","score":0.4000000059604645,"id":"https://metadata.un.org/sdg/17"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":91,"referenced_works":["https://openalex.org/W1915485278","https://openalex.org/W1932198206","https://openalex.org/W2051267297","https://openalex.org/W2060393849","https://openalex.org/W2083842231","https://openalex.org/W2108598243","https://openalex.org/W2190008860","https://openalex.org/W2194775991","https://openalex.org/W2282546112","https://openalex.org/W2530417694","https://openalex.org/W2535690855","https://openalex.org/W2535838896","https://openalex.org/W2557449848","https://openalex.org/W2560023338","https://openalex.org/W2591882872","https://openalex.org/W2630997112","https://openalex.org/W2763421725","https://openalex.org/W2767079719","https://openalex.org/W2783522756","https://openalex.org/W2783879794","https://openalex.org/W2785678896","https://openalex.org/W2789983685","https://openalex.org/W2793693263","https://openalex.org/W2796731119","https://openalex.org/W2804078698","https://openalex.org/W2893749619","https://openalex.org/W2913570153","https://openalex.org/W2917133166","https://openalex.org/W2952716587","https://openalex.org/W2962785568","https://openalex.org/W2962879692","https://openalex.org/W2962911728","https://openalex.org/W2963112338","https://openalex.org/W2963333146","https://openalex.org/W2963456518","https://openalex.org/W2963464195","https://openalex.org/W2963727650","https://openalex.org/W2963836885","https://openalex.org/W2963857746","https://openalex.org/W2964162474","https://openalex.org/W2965470910","https://openalex.org/W2970188304","https://openalex.org/W2970408908","https://openalex.org/W2970607325","https://openalex.org/W2970806306","https://openalex.org/W2982802130","https://openalex.org/W2989576588","https://openalex.org/W2989885118","https://openalex.org/W2995085126","https://openalex.org/W2997609712","https://openalex.org/W3000479830","https://openalex.org/W3005680577","https://openalex.org/W3009561768","https://openalex.org/W3034940165","https://openalex.org/W3034957837","https://openalex.org/W3035460915","https://openalex.org/W3035524453","https://openalex.org/W3035574324","https://openalex.org/W3035616549","https://openalex.org/W3038028469","https://openalex.org/W3094761124","https://openalex.org/W3105285631","https://openalex.org/W4287997702","https://openalex.org/W4288348717","https://openalex.org/W4288567374","https://openalex.org/W4295521014","https://openalex.org/W4297687186","https://openalex.org/W4297752781","https://openalex.org/W4318619660","https://openalex.org/W6695731146","https://openalex.org/W6717255582","https://openalex.org/W6728757088","https://openalex.org/W6735913928","https://openalex.org/W6745245109","https://openalex.org/W6748582592","https://openalex.org/W6752378368","https://openalex.org/W6755312952","https://openalex.org/W6759226220","https://openalex.org/W6759531683","https://openalex.org/W6762640628","https://openalex.org/W6764572792","https://openalex.org/W6764838729","https://openalex.org/W6766568290","https://openalex.org/W6767996718","https://openalex.org/W6770717842","https://openalex.org/W6771275388","https://openalex.org/W6771731465","https://openalex.org/W6773039429","https://openalex.org/W6774314701","https://openalex.org/W6774670964","https://openalex.org/W6775821767"],"related_works":["https://openalex.org/W2468279273","https://openalex.org/W2354198838","https://openalex.org/W1989130879","https://openalex.org/W2103419012","https://openalex.org/W2988126442","https://openalex.org/W1974414866","https://openalex.org/W2381850946","https://openalex.org/W4380449851","https://openalex.org/W3045696640","https://openalex.org/W2063982682"],"abstract_inverted_index":{"Training":[0],"deep":[1,217],"neural":[2],"networks":[3,112],"requires":[4],"gradient":[5],"estimation":[6],"from":[7,98,172],"data":[8,22,50,186],"batches":[9,82],"to":[10,28,85,178],"update":[11],"parameters.":[12],"Gradients":[13],"per":[14],"parameter":[15],"are":[16],"averaged":[17],"over":[18,80],"a":[19,58,63,69,99,163,193],"set":[20],"of":[21,47,183,197],"and":[23,37,219],"this":[24,89],"has":[25],"been":[26],"presumed":[27],"be":[29,86,108,206],"safe":[30],"for":[31,110,153,214],"privacy-preserving":[32],"training":[33],"in":[34],"joint,":[35],"collaborative,":[36],"federated":[38],"learning":[39],"applications.":[40],"Prior":[41],"work":[42,176],"only":[43],"showed":[44],"the":[45,184,202],"possibility":[46],"recovering":[48],"input":[49,60,75,96],"given":[51,158],"gradients":[52,79,143,191],"under":[53],"very":[54],"restrictive":[55],"conditions":[56],"\u2013":[57,103],"single":[59],"point,":[61],"or":[62,68],"network":[64],"with":[65,208],"no":[66],"non-linearities,":[67],"small":[70],"32":[71,73],"\u00d7":[72,127],"px":[74],"batch.":[76,187],"Therefore,":[77],"averaging":[78],"larger":[81,100],"was":[83],"thought":[84],"safe.":[87],"In":[88],"work,":[90],"we":[91],"introduce":[92],"GradInversion,":[93,212],"using":[94],"which":[95],"images":[97,204],"batch":[101,221],"(8":[102],"48":[104],"images)":[105],"can":[106,205],"also":[107,149],"recovered":[109,207],"large":[111,195,220],"such":[113,121,199],"as":[114,122],"ResNets":[115],"(50":[116],"layers),":[117],"on":[118],"complex":[119,215],"datasets":[120],"ImageNet":[123],"(1000":[124],"classes,":[125],"224":[126,128],"px).":[129],"We":[130,148,160,188],"formulate":[131],"an":[132,151,180],"optimization":[133],"task":[134],"that":[135,190,200],"converts":[136],"random":[137,174],"noise":[138],"into":[139],"natural":[140],"images,":[141],"matching":[142],"while":[144],"regularizing":[145],"image":[146],"fidelity.":[147],"propose":[150,162],"algorithm":[152],"target":[154],"class":[155],"label":[156],"recovery":[157],"gradients.":[159],"further":[161],"group":[164],"consistency":[165],"regularization":[166],"framework,":[167],"where":[168],"multiple":[169],"agents":[170],"starting":[171],"different":[173],"seeds":[175],"together":[177],"find":[179],"enhanced":[181],"reconstruction":[182],"original":[185],"show":[189],"encode":[192],"surprisingly":[194],"amount":[196],"information,":[198],"all":[201],"individual":[203],"high":[209],"fidelity":[210],"via":[211],"even":[213],"datasets,":[216],"networks,":[218],"sizes.":[222]},"counts_by_year":[{"year":2026,"cited_by_count":10},{"year":2025,"cited_by_count":107},{"year":2024,"cited_by_count":125},{"year":2023,"cited_by_count":97},{"year":2022,"cited_by_count":53},{"year":2021,"cited_by_count":20},{"year":2020,"cited_by_count":1}],"updated_date":"2026-04-11T08:14:18.477133","created_date":"2025-10-10T00:00:00"}