ServiceNow/PartialBROAD

---
license: cc-by-4.0
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': n01440764
'1': n01443537
'2': n01484850
'3': n01491361
'4': n01494475
'5': n01496331
'6': n01498041
'7': n01514668
'8': n01514859
'9': n01518878
'10': n01530575
'11': n01531178
'12': n01532829
'13': n01534433
'14': n01537544
'15': n01558993
'16': n01560419
'17': n01580077
'18': n01582220
'19': n01592084
'20': n01601694
'21': n01608432
'22': n01614925
'23': n01616318
'24': n01622779
'25': n01629819
'26': n01630670
'27': n01631663
'28': n01632458
'29': n01632777
'30': n01641577
'31': n01644373
'32': n01644900
'33': n01664065
'34': n01665541
'35': n01667114
'36': n01667778
'37': n01669191
'38': n01675722
'39': n01677366
'40': n01682714
'41': n01685808
'42': n01687978
'43': n01688243
'44': n01689811
'45': n01692333
'46': n01693334
'47': n01694178
'48': n01695060
'49': n01697457
'50': n01698640
'51': n01704323
'52': n01728572
'53': n01728920
'54': n01729322
'55': n01729977
'56': n01734418
'57': n01735189
'58': n01737021
'59': n01739381
'60': n01740131
'61': n01742172
'62': n01744401
'63': n01748264
'64': n01749939
'65': n01751748
'66': n01753488
'67': n01755581
'68': n01756291
'69': n01768244
'70': n01770081
'71': n01770393
'72': n01773157
'73': n01773549
'74': n01773797
'75': n01774384
'76': n01774750
'77': n01775062
'78': n01776313
'79': n01784675
'80': n01795545
'81': n01796340
'82': n01797886
'83': n01798484
'84': n01806143
'85': n01806567
'86': n01807496
'87': n01817953
'88': n01818515
'89': n01819313
'90': n01820546
'91': n01824575
'92': n01828970
'93': n01829413
'94': n01833805
'95': n01843065
'96': n01843383
'97': n01847000
'98': n01855032
'99': n01855672
'100': n01860187
'101': n01871265
'102': n01872401
'103': n01873310
'104': n01877812
'105': n01882714
'106': n01883070
'107': n01910747
'108': n01914609
'109': n01917289
'110': n01924916
'111': n01930112
'112': n01943899
'113': n01944390
'114': n01945685
'115': n01950731
'116': n01955084
'117': n01968897
'118': n01978287
'119': n01978455
'120': n01980166
'121': n01981276
'122': n01983481
'123': n01984695
'124': n01985128
'125': n01986214
'126': n01990800
'127': n02002556
'128': n02002724
'129': n02006656
'130': n02007558
'131': n02009229
'132': n02009912
'133': n02011460
'134': n02012849
'135': n02013706
'136': n02017213
'137': n02018207
'138': n02018795
'139': n02025239
'140': n02027492
'141': n02028035
'142': n02033041
'143': n02037110
'144': n02051845
'145': n02056570
'146': n02058221
'147': n02066245
'148': n02071294
'149': n02074367
'150': n02077923
'151': n02085620
'152': n02085782
'153': n02085936
'154': n02086079
'155': n02086240
'156': n02086646
'157': n02086910
'158': n02087046
'159': n02087394
'160': n02088094
'161': n02088238
'162': n02088364
'163': n02088466
'164': n02088632
'165': n02089078
'166': n02089867
'167': n02089973
'168': n02090379
'169': n02090622
'170': n02090721
'171': n02091032
'172': n02091134
'173': n02091244
'174': n02091467
'175': n02091635
'176': n02091831
'177': n02092002
'178': n02092339
'179': n02093256
'180': n02093428
'181': n02093647
'182': n02093754
'183': n02093859
'184': n02093991
'185': n02094114
'186': n02094258
'187': n02094433
'188': n02095314
'189': n02095570
'190': n02095889
'191': n02096051
'192': n02096177
'193': n02096294
'194': n02096437
'195': n02096585
'196': n02097047
'197': n02097130
'198': n02097209
'199': n02097298
'200': n02097474
'201': n02097658
'202': n02098105
'203': n02098286
'204': n02098413
'205': n02099267
'206': n02099429
'207': n02099601
'208': n02099712
'209': n02099849
'210': n02100236
'211': n02100583
'212': n02100735
'213': n02100877
'214': n02101006
'215': n02101388
'216': n02101556
'217': n02102040
'218': n02102177
'219': n02102318
'220': n02102480
'221': n02102973
'222': n02104029
'223': n02104365
'224': n02105056
'225': n02105162
'226': n02105251
'227': n02105412
'228': n02105505
'229': n02105641
'230': n02105855
'231': n02106030
'232': n02106166
'233': n02106382
'234': n02106550
'235': n02106662
'236': n02107142
'237': n02107312
'238': n02107574
'239': n02107683
'240': n02107908
'241': n02108000
'242': n02108089
'243': n02108422
'244': n02108551
'245': n02108915
'246': n02109047
'247': n02109525
'248': n02109961
'249': n02110063
'250': n02110185
'251': n02110341
'252': n02110627
'253': n02110806
'254': n02110958
'255': n02111129
'256': n02111277
'257': n02111500
'258': n02111889
'259': n02112018
'260': n02112137
'261': n02112350
'262': n02112706
'263': n02113023
'264': n02113186
'265': n02113624
'266': n02113712
'267': n02113799
'268': n02113978
'269': n02114367
'270': n02114548
'271': n02114712
'272': n02114855
'273': n02115641
'274': n02115913
'275': n02116738
'276': n02117135
'277': n02119022
'278': n02119789
'279': n02120079
'280': n02120505
'281': n02123045
'282': n02123159
'283': n02123394
'284': n02123597
'285': n02124075
'286': n02125311
'287': n02127052
'288': n02128385
'289': n02128757
'290': n02128925
'291': n02129165
'292': n02129604
'293': n02130308
'294': n02132136
'295': n02133161
'296': n02134084
'297': n02134418
'298': n02137549
'299': n02138441
'300': n02165105
'301': n02165456
'302': n02167151
'303': n02168699
'304': n02169497
'305': n02172182
'306': n02174001
'307': n02177972
'308': n02190166
'309': n02206856
'310': n02219486
'311': n02226429
'312': n02229544
'313': n02231487
'314': n02233338
'315': n02236044
'316': n02256656
'317': n02259212
'318': n02264363
'319': n02268443
'320': n02268853
'321': n02276258
'322': n02277742
'323': n02279972
'324': n02280649
'325': n02281406
'326': n02281787
'327': n02317335
'328': n02319095
'329': n02321529
'330': n02325366
'331': n02326432
'332': n02328150
'333': n02342885
'334': n02346627
'335': n02356798
'336': n02361337
'337': n02363005
'338': n02364673
'339': n02389026
'340': n02391049
'341': n02395406
'342': n02396427
'343': n02397096
'344': n02398521
'345': n02403003
'346': n02408429
'347': n02410509
'348': n02412080
'349': n02415577
'350': n02417914
'351': n02422106
'352': n02422699
'353': n02423022
'354': n02437312
'355': n02437616
'356': n02441942
'357': n02442845
'358': n02443114
'359': n02443484
'360': n02444819
'361': n02445715
'362': n02447366
'363': n02454379
'364': n02457408
'365': n02480495
'366': n02480855
'367': n02481823
'368': n02483362
'369': n02483708
'370': n02484975
'371': n02486261
'372': n02486410
'373': n02487347
'374': n02488291
'375': n02488702
'376': n02489166
'377': n02490219
'378': n02492035
'379': n02492660
'380': n02493509
'381': n02493793
'382': n02494079
'383': n02497673
'384': n02500267
'385': n02504013
'386': n02504458
'387': n02509815
'388': n02510455
'389': n02514041
'390': n02526121
'391': n02536864
'392': n02606052
'393': n02607072
'394': n02640242
'395': n02641379
'396': n02643566
'397': n02655020
'398': n02666196
'399': n02667093
'400': n02669723
'401': n02672831
'402': n02676566
'403': n02687172
'404': n02690373
'405': n02692877
'406': n02699494
'407': n02701002
'408': n02704792
'409': n02708093
'410': n02727426
'411': n02730930
'412': n02747177
'413': n02749479
'414': n02769748
'415': n02776631
'416': n02777292
'417': n02782093
'418': n02783161
'419': n02786058
'420': n02787622
'421': n02788148
'422': n02790996
'423': n02791124
'424': n02791270
'425': n02793495
'426': n02794156
'427': n02795169
'428': n02797295
'429': n02799071
'430': n02802426
'431': n02804414
'432': n02804610
'433': n02807133
'434': n02808304
'435': n02808440
'436': n02814533
'437': n02814860
'438': n02815834
'439': n02817516
'440': n02823428
'441': n02823750
'442': n02825657
'443': n02834397
'444': n02835271
'445': n02837789
'446': n02840245
'447': n02841315
'448': n02843684
'449': n02859443
'450': n02860847
'451': n02865351
'452': n02869837
'453': n02870880
'454': n02871525
'455': n02877765
'456': n02879718
'457': n02883205
'458': n02892201
'459': n02892767
'460': n02894605
'461': n02895154
'462': n02906734
'463': n02909870
'464': n02910353
'465': n02916936
'466': n02917067
'467': n02927161
'468': n02930766
'469': n02939185
'470': n02948072
'471': n02950826
'472': n02951358
'473': n02951585
'474': n02963159
'475': n02965783
'476': n02966193
'477': n02966687
'478': n02971356
'479': n02974003
'480': n02977058
'481': n02978881
'482': n02979186
'483': n02980441
'484': n02981792
'485': n02988304
'486': n02992211
'487': n02992529
'488': n02999410
'489': n03000134
'490': n03000247
'491': n03000684
'492': n03014705
'493': n03016953
'494': n03017168
'495': n03018349
'496': n03026506
'497': n03028079
'498': n03032252
'499': n03041632
'500': n03042490
'501': n03045698
'502': n03047690
'503': n03062245
'504': n03063599
'505': n03063689
'506': n03065424
'507': n03075370
'508': n03085013
'509': n03089624
'510': n03095699
'511': n03100240
'512': n03109150
'513': n03110669
'514': n03124043
'515': n03124170
'516': n03125729
'517': n03126707
'518': n03127747
'519': n03127925
'520': n03131574
'521': n03133878
'522': n03134739
'523': n03141823
'524': n03146219
'525': n03160309
'526': n03179701
'527': n03180011
'528': n03187595
'529': n03188531
'530': n03196217
'531': n03197337
'532': n03201208
'533': n03207743
'534': n03207941
'535': n03208938
'536': n03216828
'537': n03218198
'538': n03220513
'539': n03223299
'540': n03240683
'541': n03249569
'542': n03250847
'543': n03255030
'544': n03259280
'545': n03271574
'546': n03272010
'547': n03272562
'548': n03290653
'549': n03291819
'550': n03297495
'551': n03314780
'552': n03325584
'553': n03337140
'554': n03344393
'555': n03345487
'556': n03347037
'557': n03355925
'558': n03372029
'559': n03376595
'560': n03379051
'561': n03384352
'562': n03388043
'563': n03388183
'564': n03388549
'565': n03393912
'566': n03394916
'567': n03400231
'568': n03404251
'569': n03417042
'570': n03424325
'571': n03425413
'572': n03443371
'573': n03444034
'574': n03445777
'575': n03445924
'576': n03447447
'577': n03447721
'578': n03450230
'579': n03452741
'580': n03457902
'581': n03459775
'582': n03461385
'583': n03467068
'584': n03476684
'585': n03476991
'586': n03478589
'587': n03481172
'588': n03482405
'589': n03483316
'590': n03485407
'591': n03485794
'592': n03492542
'593': n03494278
'594': n03495258
'595': n03496892
'596': n03498962
'597': n03527444
'598': n03529860
'599': n03530642
'600': n03532672
'601': n03534580
'602': n03535780
'603': n03538406
'604': n03544143
'605': n03584254
'606': n03584829
'607': n03590841
'608': n03594734
'609': n03594945
'610': n03595614
'611': n03598930
'612': n03599486
'613': n03602883
'614': n03617480
'615': n03623198
'616': n03627232
'617': n03630383
'618': n03633091
'619': n03637318
'620': n03642806
'621': n03649909
'622': n03657121
'623': n03658185
'624': n03661043
'625': n03662601
'626': n03666591
'627': n03670208
'628': n03673027
'629': n03676483
'630': n03680355
'631': n03690938
'632': n03691459
'633': n03692522
'634': n03697007
'635': n03706229
'636': n03709823
'637': n03710193
'638': n03710637
'639': n03710721
'640': n03717622
'641': n03720891
'642': n03721384
'643': n03724870
'644': n03729826
'645': n03733131
'646': n03733281
'647': n03733805
'648': n03742115
'649': n03743016
'650': n03759954
'651': n03761084
'652': n03763968
'653': n03764736
'654': n03769881
'655': n03770439
'656': n03770679
'657': n03773504
'658': n03775071
'659': n03775546
'660': n03776460
'661': n03777568
'662': n03777754
'663': n03781244
'664': n03782006
'665': n03785016
'666': n03786901
'667': n03787032
'668': n03788195
'669': n03788365
'670': n03791053
'671': n03792782
'672': n03792972
'673': n03793489
'674': n03794056
'675': n03796401
'676': n03803284
'677': n03804744
'678': n03814639
'679': n03814906
'680': n03825788
'681': n03832673
'682': n03837869
'683': n03838899
'684': n03840681
'685': n03841143
'686': n03843555
'687': n03854065
'688': n03857828
'689': n03866082
'690': n03868242
'691': n03868863
'692': n03871628
'693': n03873416
'694': n03874293
'695': n03874599
'696': n03876231
'697': n03877472
'698': n03877845
'699': n03884397
'700': n03887697
'701': n03888257
'702': n03888605
'703': n03891251
'704': n03891332
'705': n03895866
'706': n03899768
'707': n03902125
'708': n03903868
'709': n03908618
'710': n03908714
'711': n03916031
'712': n03920288
'713': n03924679
'714': n03929660
'715': n03929855
'716': n03930313
'717': n03930630
'718': n03933933
'719': n03935335
'720': n03937543
'721': n03938244
'722': n03942813
'723': n03944341
'724': n03947888
'725': n03950228
'726': n03954731
'727': n03956157
'728': n03958227
'729': n03961711
'730': n03967562
'731': n03970156
'732': n03976467
'733': n03976657
'734': n03977966
'735': n03980874
'736': n03982430
'737': n03983396
'738': n03991062
'739': n03992509
'740': n03995372
'741': n03998194
'742': n04004767
'743': n04005630
'744': n04008634
'745': n04009552
'746': n04019541
'747': n04023962
'748': n04026417
'749': n04033901
'750': n04033995
'751': n04037443
'752': n04039381
'753': n04040759
'754': n04041544
'755': n04044716
'756': n04049303
'757': n04065272
'758': n04067472
'759': n04069434
'760': n04070727
'761': n04074963
'762': n04081281
'763': n04086273
'764': n04090263
'765': n04099969
'766': n04111531
'767': n04116512
'768': n04118538
'769': n04118776
'770': n04120489
'771': n04125021
'772': n04127249
'773': n04131690
'774': n04133789
'775': n04136333
'776': n04141076
'777': n04141327
'778': n04141975
'779': n04146614
'780': n04147183
'781': n04149813
'782': n04152593
'783': n04153751
'784': n04154565
'785': n04162706
'786': n04179913
'787': n04192698
'788': n04200800
'789': n04201297
'790': n04204238
'791': n04204347
'792': n04208210
'793': n04209133
'794': n04209239
'795': n04228054
'796': n04229816
'797': n04235860
'798': n04238763
'799': n04239074
'800': n04243546
'801': n04251144
'802': n04252077
'803': n04252225
'804': n04254120
'805': n04254680
'806': n04254777
'807': n04258138
'808': n04259630
'809': n04263257
'810': n04264628
'811': n04265275
'812': n04266014
'813': n04270147
'814': n04273569
'815': n04275548
'816': n04277352
'817': n04285008
'818': n04286575
'819': n04296562
'820': n04310018
'821': n04311004
'822': n04311174
'823': n04317175
'824': n04325704
'825': n04326547
'826': n04328186
'827': n04330267
'828': n04332243
'829': n04335435
'830': n04336792
'831': n04344873
'832': n04346328
'833': n04347754
'834': n04350905
'835': n04355338
'836': n04355933
'837': n04356056
'838': n04357314
'839': n04366367
'840': n04367480
'841': n04370456
'842': n04371430
'843': n04371774
'844': n04372370
'845': n04376876
'846': n04380533
'847': n04389033
'848': n04392985
'849': n04398044
'850': n04399382
'851': n04404412
'852': n04409515
'853': n04417672
'854': n04418357
'855': n04423845
'856': n04428191
'857': n04429376
'858': n04435653
'859': n04442312
'860': n04443257
'861': n04447861
'862': n04456115
'863': n04458633
'864': n04461696
'865': n04462240
'866': n04465501
'867': n04467665
'868': n04476259
'869': n04479046
'870': n04482393
'871': n04483307
'872': n04485082
'873': n04486054
'874': n04487081
'875': n04487394
'876': n04493381
'877': n04501370
'878': n04505470
'879': n04507155
'880': n04509417
'881': n04515003
'882': n04517823
'883': n04522168
'884': n04523525
'885': n04525038
'886': n04525305
'887': n04532106
'888': n04532670
'889': n04536866
'890': n04540053
'891': n04542943
'892': n04548280
'893': n04548362
'894': n04550184
'895': n04552348
'896': n04553703
'897': n04554684
'898': n04557648
'899': n04560804
'900': n04562935
'901': n04579145
'902': n04579432
'903': n04584207
'904': n04589890
'905': n04590129
'906': n04591157
'907': n04591713
'908': n04592741
'909': n04596742
'910': n04597913
'911': n04599235
'912': n04604644
'913': n04606251
'914': n04612504
'915': n04613696
'916': n06359193
'917': n06596364
'918': n06785654
'919': n06794110
'920': n06874185
'921': n07248320
'922': n07565083
'923': n07579787
'924': n07583066
'925': n07584110
'926': n07590611
'927': n07613480
'928': n07614500
'929': n07615774
'930': n07684084
'931': n07693725
'932': n07695742
'933': n07697313
'934': n07697537
'935': n07711569
'936': n07714571
'937': n07714990
'938': n07715103
'939': n07716358
'940': n07716906
'941': n07717410
'942': n07717556
'943': n07718472
'944': n07718747
'945': n07720875
'946': n07730033
'947': n07734744
'948': n07742313
'949': n07745940
'950': n07747607
'951': n07749582
'952': n07753113
'953': n07753275
'954': n07753592
'955': n07754684
'956': n07760859
'957': n07768694
'958': n07802026
'959': n07831146
'960': n07836838
'961': n07860988
'962': n07871810
'963': n07873807
'964': n07875152
'965': n07880968
'966': n07892512
'967': n07920052
'968': n07930864
'969': n07932039
'970': n09193705
'971': n09229709
'972': n09246464
'973': n09256479
'974': n09288635
'975': n09332890
'976': n09399592
'977': n09421951
'978': n09428293
'979': n09468604
'980': n09472597
'981': n09835506
'982': n10148035
'983': n10565667
'984': n11879895
'985': n11939491
'986': n12057211
'987': n12144580
'988': n12267677
'989': n12620546
'990': n12768682
'991': n12985857
'992': n12998815
'993': n13037406
'994': n13040303
'995': n13044778
'996': n13052670
'997': n13054560
'998': n13133613
'999': n15075141
- name: original_filename
dtype: string
- name: original_hash
dtype: string
splits:
- name: synthetic_gan
num_bytes: 241800515.64
num_examples: 24999
- name: synthetic_diffusion
num_bytes: 329296506.0
num_examples: 25000
- name: adversarial_autoattack_resnet
num_bytes: 755454273.0
num_examples: 5000
- name: adversarial_autoattack_vit
num_bytes: 2217501074.0
num_examples: 5000
- name: adversarial_pgd_resnet
num_bytes: 755454474.0
num_examples: 5000
- name: adversarial_pgd_vit
num_bytes: 2217501084.0
num_examples: 5000
download_size: 6513646801
dataset_size: 6517007926.639999
pretty_name: BROAD
size_categories:
- 10K<n<100K
tags:
- imagenet
- OOD detection
- distribution shift
---
# Partial dataset used to build BROAD (Benchmarking Resilience Over Anomaly Diversity )

Refer to [this repo ](https://github.com/ServiceNow/broad) to build the complete BROAD dataset.

The partial data included here contains synthetica images from BROAD and encoded unrecognizable images given by adversarial perturbations of imagenet samples. Decoding is implemented in the repo referred above.

## Dataset Description

The BROAD dataset was introduced to benchmark OOD detection methods against a broader variety of distribution shifts in the paper
Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection.

Each split of BROAD is designed to be close (but different) to the [ImageNet](https://www.image-net.org/index.php) distribution.

### Dataset Summary

BROAD is comprised of 16 splits, 9 of which can be downloaded from this page. The remaining 7 can be obtained through external links.
We first describe the splits available from this hub, and then specify the external splits and how to get them. Please refer to Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection for more detailed description of the data and its acquisition.

### Included Splits

- **Clean** is comprised of 36157 images from the original validation set of ILSVRC2012. They are used as in-distribution in BROAD.
- **Adversarial Autoattack Resnet**, **Adversarial Autoattack ViT**, **Adversarial PGD Resnet** and **Adversarial PGD ViT** are splits each comprised of 5,000 adversarial perturbations of clean validation images, using a perturbation budget of 0.05 with the L-infinity norm. These attacks are computed against a trained ResNet-50 and a trained ViT-b/16. PGD uses 40 iterations and for Autoattack, only the attack model achieving the most confident misclassification is kept.
- **Synthetic Gan** and **Synthetic Diffusion** are each comprised of 25,000 synthetic images generated to imitate the ImageNet distribution. For Synthetic Gan, a conditional BigGan architecture was used to generate 25 artificial samples from each ImageNet class. For Synthetic diffusion, we leveraged stable diffusion models to generate 25 artificial samples per class using the prompt "High quality image of a {class_name}".
- **CoComageNet** is a novel split from the [CoCo](https://cocodataset.org/#home) dataset comprised of 2000 images, each featuring multiple distinct classes of ImageNet. Each image of CoComageNet thus features multiple objects, at least two of which have distinct ImageNet labels. More details on the construction of CoComageNet can be found in the paper.
- **CoComageNet-mono** is built similarly to CoComageNet, except each image only has one object with ImageNet label. It is designed as an ablation, to isolate the effect of having instances of multiple labels from other distributional shifts in CoComageNet.

### External Splits

- **iNaturalist** is a split of the original [iNaturalist2017 dataset](https://github.com/visipedia/inat_comp/tree/master/2017) designed for OOD detection with ImageNet as in-distribution. It was introduced in [MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Space](https://arxiv.org/pdf/2105.01879.pdf) and can be downloaded [here](http://pages.cs.wisc.edu/~huangrui/imagenet_ood_dataset/iNaturalist.tar.gz).
- **ImageNet-O** was introduced in [Natural Adversarial Examples](https://arxiv.org/pdf/1907.07174.pdf) and is comprised of natural examples that were selected for their high classification confidence by CNNs. It can be downloaded [here](https://people.eecs.berkeley.edu/~hendrycks/imagenet-o.tar).
- **OpenImage-O** is a subset of the OpenImage dataset that was built similarly to ImageNet-O in [ViM: Out-Of-Distribution with Virtual-logit Matching](https://arxiv.org/pdf/2203.10807.pdf). The file list can be accessed [here](https://github.com/haoqiwang/vim/blob/master/datalists/openimage_o.txt).
- **Defocus blur**, **Gaussian noise**, **Snow** and **Brightness** are all existing splits of the [ImageNet-C dataset](https://github.com/hendrycks/robustness). For BROAD, only the highest strength of corruption (5/5) is used.

### LICENSE

This work is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by/4.0/deed.en_US">Creative Commons Attribution 4.0 Unported License</a>.

许可证：CC BY 4.0
数据集信息：
特征：
- 名称：image
数据类型：图像
- 名称：label
数据类型：分类标签，共包含1000个类别，编号0至999分别对应ImageNet的WordNet类标识符n01440764至n15075141
- 名称：original_filename
数据类型：字符串
- 名称：original_hash
数据类型：字符串
拆分：
- 名称：synthetic_gan（合成GAN）
字节大小：241800515.64
样本数量：24999
- 名称：synthetic_diffusion（合成扩散）
字节大小：329296506.0
样本数量：25000
- 名称：adversarial_autoattack_resnet（对抗自动攻击残差网络（ResNet））
字节大小：755454273.0
样本数量：5000
- 名称：adversarial_autoattack_vit（对抗自动攻击视觉Transformer（ViT））
字节大小：2217501074.0
样本数量：5000
- 名称：adversarial_pgd_resnet（对抗PGD（投影梯度下降）残差网络（ResNet））
字节大小：755454474.0
样本数量：5000
- 名称：adversarial_pgd_vit（对抗PGD（投影梯度下降）视觉Transformer（ViT））
字节大小：2217501084.0
样本数量：5000
下载总大小：6513646801 字节
数据集总大小：6517007926.64 字节
可读名称：BROAD
样本量区间：10000 < 样本数 < 100000
标签：
- ImageNet
- 分布外检测（Out-of-Distribution Detection，OOD检测）
- 分布偏移

---
# 用于构建BROAD（基准测试异常多样性下的鲁棒性）的部分数据集
参考[此仓库](https://github.com/ServiceNow/broad)以构建完整的BROAD数据集。
本部分数据包含BROAD中的合成图像，以及由ImageNet样本经对抗扰动生成的不可识别编码图像。解码代码可在上述引用的仓库中获取。

## 数据集描述
BROAD数据集由论文《Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection》提出，用于基准测试分布外（OOD）检测方法，以应对更广泛的分布偏移场景。

BROAD的每个拆分集均设计为与[ImageNet](https://www.image-net.org/index.php)分布相近但存在差异。

### 数据集摘要
BROAD共包含16个拆分集，其中9个可在此页面下载，剩余7个可通过外部链接获取。下文将先介绍本页面提供的拆分集，随后说明外部拆分集及其获取方式。有关数据及其获取方式的详细描述，请参阅论文《Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection》。

### 本页面包含的拆分集
- **Clean（干净样本）**：包含来自ILSVRC2012原始验证集的36157张图像，作为BROAD中的分布内样本。
- **Adversarial Autoattack Resnet（对抗自动攻击残差网络（ResNet））**、**Adversarial Autoattack ViT（对抗自动攻击视觉Transformer（ViT））**、**Adversarial PGD Resnet（对抗PGD残差网络（ResNet））**和**Adversarial PGD ViT（对抗PGD视觉Transformer（ViT））**：每个拆分集均包含5000张对干净验证图像的对抗扰动样本，扰动预算采用L∞范数下的0.05。这些攻击均针对训练好的ResNet-50和训练好的ViT-b/16模型生成。PGD攻击采用40次迭代，Autoattack仅保留实现最高置信度误分类的攻击结果。
- **Synthetic Gan（合成GAN）**与**Synthetic Diffusion（合成扩散）**：每个拆分集均包含25000张用于模仿ImageNet分布的合成图像。其中合成GAN拆分采用条件BigGAN架构，为每个ImageNet类别生成25张人工样本；合成扩散拆分则借助Stable Diffusion模型，通过提示词"High quality image of a {class_name}"为每个类别生成25张人工样本。
- **CoComageNet**：源自[COCO（通用物体检测数据集）](https://cocodataset.org/#home)的新型拆分集，包含2000张图像，每张图像均包含多个不同的ImageNet类别对象。因此，CoComageNet的每张图像均包含至少两个具有不同ImageNet标签的物体。有关CoComageNet构建方式的更多细节可参阅论文。
- **CoComageNet-mono**：构建方式与CoComageNet类似，但每张图像仅包含一个带有ImageNet标签的对象。该拆分集被设计为消融实验集，用于隔离CoComageNet中多标签实例与其他分布偏移的影响。

### 外部拆分集
- **iNaturalist**：源自原始[iNaturalist2017数据集](https://github.com/visipedia/inat_comp/tree/master/2017)的拆分集，用于以ImageNet作为分布内样本的OOD检测任务。该拆分集首次出现在论文《MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Space》中，可从[此处](http://pages.cs.wisc.edu/~huangrui/imagenet_ood_dataset/iNaturalist.tar.gz)下载。
- **ImageNet-O**：首次出现在论文《Natural Adversarial Examples》中，包含经CNN分类器判定为高置信度的自然对抗样本。可从[此处](https://people.eecs.berkeley.edu/~hendrycks/imagenet-o.tar)下载。
- **OpenImage-O**：源自OpenImage数据集的子集，构建方式与ImageNet-O类似，出自论文《ViM: Out-Of-Distribution with Virtual-logit Matching》。其文件列表可从[此处](https://github.com/haoqiwang/vim/blob/master/datalists/openimage_o.txt)获取。
- **Defocus blur（散焦模糊）**、**Gaussian noise（高斯噪声）**、**Snow（降雪）**和**Brightness（亮度变化）**均为[ImageNet-C数据集](https://github.com/hendrycks/robustness)的现有拆分集。在BROAD中，仅使用扰动强度为5/5的版本。

### 许可证
本作品采用<a rel="license" href="http://creativecommons.org/licenses/by/4.0/deed.en_US">知识共享署名4.0国际许可协议</a>进行许可。

原始信息汇总

数据集概述

数据集名称

• BROAD

数据集特征

• date_captured: 数据类型为字符串
• coco_url: 数据类型为字符串
• license_name: 数据类型为字符串
• license_url: 数据类型为字符串
• coco_id: 数据类型为字符串
• image: 数据类型为图像
• label: 数据类型为int64
• flickr_url: 数据类型为字符串

数据集分割

• Clean: 包含36,157个示例，总字节数为333,801,279.747
• CocomageNet: 包含2,000个示例，总字节数为306,403,223
• CocomageNet_mono: 包含2,000个示例，总字节数为18,956,338
• Synthetic_gan: 包含24,999个示例，总字节数为242,598,071.454
• Synthetic_diffusion: 包含25,000个示例，总字节数为283,705,025
• Adversarial_autoattack_resnet: 包含5,000个示例，总字节数为40,058,245
• Adversarial_autoattack_vit: 包含5,000个示例，总字节数为35,610,460
• Adversarial_pgd_resnet: 包含5,000个示例，总字节数为65,806,170
• Adversarial_pgd_vit: 包含5,000个示例，总字节数为51,803,590

数据集大小

• 下载大小: 1,432,934,722字节
• 数据集大小: 1,378,742,402.201字节

数据集标签

• imagenet
• OOD detection
• distribution shift

数据集许可

• Creative Commons Attribution 4.0 Unported License