NAS-Bench-101:TowardsReproducibleNeuralArchitectureSearchChrisYing1AaronKlein2EstebanReal1EricChristiansen1KevinMurphy1FrankHutter2Abstractturesasanoptimizationproblem(Bakeretal.,2017;Zoph&Le,2016;...
BayesNAS:ABayesianApproachforNeuralArchitectureSearchHongpengZhou1MinghaoYang1JunWang2WeiPan1Abstractetal.,2018;Caietal.,2018)methodstakethegenerationofaneuralArchitectureasanagent’sactionwiththea...
Path-LevelNetworkTransformationforEfficientArchitectureSearchHanCai1JiachengYang1WeinanZhang1SongHan2YongYu1Abstractthisprocesstypicallyrequiresyearsofextensiveinvestiga-tionbyhumanexperts,whichisn...
EfficientNeuralArchitectureSearchviaParameterSharingHieuPham12MelodyY.Guan3BarretZoph1QuocV.Le1JeffDean1Abstractconsuming,e.g.Zophetal.(2018)use450GPUsfor3-4days(i.e.32,400-43,200GPUhours).Meanwhil...
InputSwitchedAffineNetworks:AnRNNArchitectureDesignedforInterpretabilityJakobN.Foerster1JustinGilmer2JaschaSohl-Dickstein3JanChorowski4DavidSussillo3AbstractquiringinterpretableMLincludeself-drivin...