作者：floodsun编译：ronghuaiyang导读
这个路线图是根据下面几个规则构建的：
从概要到细节
从老的到最新的业界领先
从通用的到细分领域的
聚焦业界领先的
你会发现很多论文很新，但是确实值得一读。
我会继续在这个路线图上添加论文。
1深度学习历史和基础1.0书
[0]Bengio,Yoshua,,andAaronCourville."Deeplearning."AnMITPressbook.(2015).[html](深度学习的圣经，你读这本书的同时可以读下面的论文.)⭐️⭐️⭐️⭐️⭐️
1.1综述
[1]LeCun,Yann,YoshuaBengio,andGeoffreyHinton."Deeplearning."(2015):436-444.[pdf](三巨头的综述)⭐️⭐️⭐️⭐️⭐️
1.2深度置信网络(DBN)(深度学习黎明之前的里程碑)
[2]Hinton,GeoffreyE.,SimonOsindero,andYee-WhyeTeh."Afastlearningalgorithmfordeepbeliefnets."(2006):1527-1554.[pdf](深度学习黎明之前)⭐️⭐️⭐️
[3]Hinton,GeoffreyE.,"Reducingthedimensionalityofdatawithneuralnetworks."(2006):504-507.[pdf](里程碑,展示了深度学习的前途)⭐️⭐️⭐️
1.3ImageNet的演进（深度学习从这里突破）
[4]Krizhevsky,Alex,IlyaSutskever,"Imagenetclassificationwithdeepconvolutionalneuralnetworks."[pdf](AlexNet,深度学习的突破)⭐️⭐️⭐️⭐️⭐️
[5]Simonyan,Karen,andAndrewZisserman."Verydeepconvolutionalnetworksforlarge-scaleimagerecognition."arXivpreprintarXiv:1409.1556(2014).[pdf](VGGNet,神经网络开始变得很深!)⭐️⭐️⭐️
[6]Szegedy,Christian,etal."Goingdeeperwithconvolutions."Proc[pdf](GoogLeNet)⭐️⭐️⭐️
[7]He,Kaiming,etal."Deepresiduallearningforimagerecognition."arXivpreprintarXiv:1512.03385(2015).[pdf](ResNet,非常非常深的神经网络,CVPR最佳论文)⭐️⭐️⭐️⭐️⭐️
1.4语音识别的演进
[8]Hinton,Geoffrey,etal."Deepneuralnetworksforacousticmodelinginspeechrecognition:Thesharedviewsoffourresearchgroups."(2012):82-97.[pdf](语音识别的突破)⭐️⭐️⭐️⭐️
[9]Graves,Alex,Abdel-rahmanMohamed,andGeoffreyHinton."Speechrecognitionwithdeeprecurrentneuralnetworks."2013IEEEinternationalconferenceonacoustics,,2013.[pdf](RNN，循环神经网络)⭐️⭐️⭐️
[10]Graves,Alex,andNavdeepJaitly."Towards-To-SpeechRecognitionwithRecurrentNeuralNetworks."[pdf]⭐️⭐️⭐️
[11]Sak,Haşim,etal."Fastandaccuraterecurrentneuralnetworkacousticmodelsforspeechrecognition."arXivpreprintarXiv:1507.06947(2015).[pdf](Google的语音识别系统)⭐️⭐️⭐️
[12]Amodei,Dario,etal."Deepspeech2:-to-speechrecognitioninenglishandmandarin."arXivpreprintarXiv:1512.02595(2015).[pdf](Baidu的语音识别系统)⭐️⭐️⭐️⭐️
[13],,,,,,,"AchievingHumanParityinConversationalSpeechRecognition."arXivpreprintarXiv:1610.05256(2016).[pdf](state-of-the-art的语音识别,Microsoft)⭐️⭐️⭐️⭐️
2深度学习方法2.1模型
[14]Hinton,GeoffreyE.,etal."Improvingneuralnetworksbypreventingco-adaptationoffeaturedetectors."arXivpreprintarXiv:1207.0580(2012).[pdf](Dropout)⭐️⭐️⭐️
[15]Srivastava,Nitish,etal."Dropout:asimplewaytopreventneuralnetworksfromoverfitting."(2014):1929-1958.[pdf]⭐️⭐️⭐️
[16]Ioffe,Sergey,andChristianSzegedy."Batchnormalization:Acceleratingdeepnetworktrainingbyreducinginternalcovariateshift."arXivpreprintarXiv:1502.03167(2015).[pdf](2015年的杰出工作)⭐️⭐️⭐️⭐️
[17]Ba,JimmyLei,JamieRyanKiros,"Layernormalization."arXivpreprintarXiv:1607.06450(2016).[pdf](BatchNormalization的升级版)⭐️⭐️⭐️⭐️
[18]Courbariaux,Matthieu,etal."BinarizedNeuralNetworks:TrainingNeuralNetworkswithWeightsandActivationsConstrainedto+1or−1."[pdf](新模型，速度快)⭐️⭐️⭐️
[19]Jaderberg,Max,etal."Decoupledneuralinterfacesusingsyntheticgradients."arXivpreprintarXiv:1608.05343(2016).[pdf](训练方法的创新，了不起的工作)⭐️⭐️⭐️⭐️⭐️
[20]Chen,Tianqi,IanGoodfellow,andJonathonShlens."Net2net:Acceleratinglearningviaknowledgetransfer."arXivpreprintarXiv:1511.05641(2015).[pdf](通过修改之前的训练过的网络来减少训练的epoch数量)⭐️⭐️⭐️
[21]Wei,Tao,etal."NetworkMorphism."arXivpreprintarXiv:1603.01670(2016).[pdf](通过修改之前的训练过的网络来减少训练的epoch数量)⭐️⭐️⭐️
2.2优化方法
[22]Sutskever,Ilya,etal."Ontheimportanceofinitializationandmomentumindeeplearning."ICML(3)28(2013):1139-1147.[pdf](动量优化器)⭐️⭐️
[23]Kingma,Diederik,andJimmyBa."Adam:Amethodforstochasticoptimization."arXivpreprintarXiv:1412.6980(2014).[pdf](可能是最常用的优化器)⭐️⭐️⭐️
[24]Andrychowicz,Marcin,etal."Learningtolearnbygradientdescentbygradientdescent."arXivpreprintarXiv:1606.04474(2016).[pdf](神经元优化器，杰出的工作)⭐️⭐️⭐️⭐️⭐️
[25]Han,Song,HuiziMao,"Deepcompression:Compressingdeepneuralnetworkwithpruning,trainedquantizationandhuffmancoding."CoRR,abs/1510.001492(2015).[pdf](ICLR最佳论文,是神经网络运行更快的新方向)⭐️⭐️⭐️⭐️⭐️
[26]Iandola,ForrestN.,etal."SqueezeNet:AlexNet-levelaccuracywith50xfewerparametersand1MBmodelsize."arXivpreprintarXiv:1602.07360(2016).[pdf](也是一个优化神经网络的新方向)⭐️⭐️⭐️⭐️
2.3非监督学习/深度生成模型
谷歌[27]Le,QuocV."Buildinghigh-levelfeaturesusinglargescaleunsupervisedlearning."2013IEEEinternationalconferenceonacoustics,,2013.[pdf](里程碑,吴恩达,谷歌大脑项目,认猫)⭐️⭐️⭐️⭐️
[28]Kingma,DiederikP.,andMaxWelling."Auto-encodingvariationalbayes."arXivpreprintarXiv:1312.6114(2013).[pdf](VAE)⭐️⭐️⭐️⭐️
[29]Goodfellow,Ian,etal."Generativeadversarialnets."[pdf](生成对抗网络，超酷的点子)⭐️⭐️⭐️⭐️⭐️
[30]Radford,Alec,LukeMetz,andSoumithChintala."Unsupervisedrepresentationlearningwithdeepconvolutionalgenerativeadversarialnetworks."arXivpreprintarXiv:1511.06434(2015).[pdf](DCGAN)⭐️⭐️⭐️⭐️
[31]Gregor,Karol,etal."DRAW:Arecurrentneuralnetworkforimagegeneration."arXivpreprintarXiv:1502.04623(2015).[pdf](带注意力机制的VAE,出色的工作)⭐️⭐️⭐️⭐️⭐️
[32]Oord,Aaronvanden,NalKalchbrenner,andKorayKavukcuoglu."Pixelrecurrentneuralnetworks."arXivpreprintarXiv:1601.06759(2016).[pdf](PixelRNN)⭐️⭐️⭐️⭐️
[33]Oord,Aaronvanden,etal."ConditionalimagegenerationwithPixelCNNdecoders."arXivpreprintarXiv:1606.05328(2016).[pdf](PixelCNN)⭐️⭐️⭐️⭐️
2.4循环神经网络（RNN）/序列到序列（Seq-to-Seq）的模型
[34]Graves,Alex."Generatingsequenceswithrecurrentneuralnetworks."arXivpreprintarXiv:1308.0850(2013).[pdf](LSTM,非常好的生成结果,展示了RNN的威力)⭐️⭐️⭐️⭐️
[35]Cho,Kyunghyun,etal."LearningphraserepresentationsusingRNNencoder-decoderforstatisticalmachinetranslation."arXivpreprintarXiv:1406.1078(2014).[pdf](第一个Seq-to-Seq的论文)⭐️⭐️⭐️⭐️
[36]Sutskever,Ilya,OriolVinyals,"Sequencetosequencelearningwithneuralnetworks."[pdf](杰出的工作)⭐️⭐️⭐️⭐️⭐️
[37]Bahdanau,Dzmitry,KyungHyunCho,andYoshuaBengio."NeuralMachineTranslationbyJointlyLearningtoAlignandTranslate."arXivpreprintarXiv:1409.0473(2014).[pdf]⭐️⭐️⭐️⭐️
[38]Vinyals,Oriol,andQuocLe."Aneuralconversationalmodel."arXivpreprintarXiv:1506.05869(2015).[pdf](Seq-to-Seq在聊天机器人上应用)⭐️⭐️⭐️
2.5神经元图灵机
[39]Graves,Alex,GregWayne,andIvoDanihelka."Neuralturingmachines."arXivpreprintarXiv:1410.5401(2014).[pdf](未来计算机的基本原型)⭐️⭐️⭐️⭐️⭐️
[40]Zaremba,Wojciech,andIlyaSutskever."ReinforcementlearningneuralTuringmachines."arXivpreprintarXiv:1505.00521362(2015).[pdf]⭐️⭐️⭐️
[41]Weston,Jason,SumitChopra,andAntoineBordes."Memorynetworks."arXivpreprintarXiv:1410.3916(2014).[pdf]⭐️⭐️⭐️
[42]Sukhbaatar,Sainbayar,JasonWeston,andRobFergus."-to-memorynetworks."[pdf]⭐️⭐️⭐️⭐️
[43]Vinyals,Oriol,MeireFortunato,andNavdeepJaitly."Pointernetworks."[pdf]⭐️⭐️⭐️⭐️
[44]Graves,Alex,etal."Hybridcomputingusinganeuralnetworkwithdynamicexternalmemory."Nature(2016).[pdf](里程碑，综合了上面这些论文的思想)⭐️⭐️⭐️⭐️⭐️
2.6深度强化学习
[45]Mnih,Volodymyr,etal."Playingatariwithdeepreinforcementlearning."arXivpreprintarXiv:1312.5602(2013).[pdf])(第一篇命名为深度强化学习的论文)⭐️⭐️⭐️⭐️
[46]Mnih,Volodymyr,etal."Human-levelcontrolthroughdeepreinforcementlearning."(2015):529-533.[pdf](里程碑)⭐️⭐️⭐️⭐️⭐️
[47]Wang,Ziyu,NandodeFreitas,andMarcLanctot."Duelingnetworkarchitecturesfordeepreinforcementlearning."arXivpreprintarXiv:1511.06581(2015).[pdf](ICLR最佳论文，非常好的点子)⭐️⭐️⭐️⭐️
[48]Mnih,Volodymyr,etal."Asynchronousmethodsfordeepreinforcementlearning."arXivpreprintarXiv:1602.01783(2016).[pdf](业界领先的方法)⭐️⭐️⭐️⭐️⭐️
[49]Lillicrap,TimothyP.,etal."Continuouscontrolwithdeepreinforcementlearning."arXivpreprintarXiv:1509.02971(2015).[pdf](DDPG)⭐️⭐️⭐️⭐️
[50]Gu,Shixiang,etal."ContinuousDeepQ-LearningwithModel-basedAcceleration."arXivpreprintarXiv:1603.00748(2016).[pdf](NAF)⭐️⭐️⭐️⭐️
[51]Schulman,John,etal."Trustregionpolicyoptimization."CoRR,abs/1502.05477(2015).[pdf](TRPO)⭐️⭐️⭐️⭐️
[52]Silver,David,etal."MasteringthegameofGowithdeepneuralnetworksandtreesearch."(2016):484-489.[pdf](阿尔法狗)⭐️⭐️⭐️⭐️⭐️
2.7深度迁移学习/终身学习/特别针对是强化学习
[53]Bengio,Yoshua."DeepLearningofRepresentationsforUnsupervisedandTransferLearning."ICMLUnsupervisedandTransferLearning27(2012):17-36.[pdf](一个教程)⭐️⭐️⭐️
[54]Silver,DanielL.,QiangYang,andLianghaoLi."LifelongMachineLearningSystems:BeyondLearningAlgorithms."AAAISpringSymposium:[pdf](一个关于终身学习的简单的探讨)⭐️⭐️⭐️
[55]Hinton,Geoffrey,OriolVinyals,andJeffDean."Distillingtheknowledgeinaneuralnetwork."arXivpreprintarXiv:1503.02531(2015).[pdf](深度学习教父的工作)⭐️⭐️⭐️⭐️
[56]Rusu,AndreiA.,etal."Policydistillation."arXivpreprintarXiv:1511.06295(2015).[pdf](强化学习的领域)⭐️⭐️⭐️
[57]Parisotto,Emilio,JimmyLeiBa,andRuslanSalakhutdinov."Actor-mimic:Deepmultitaskandtransferreinforcementlearning."arXivpreprintarXiv:1511.06342(2015).[pdf](强化学习的领域)⭐️⭐️⭐️
[58]Rusu,AndreiA.,etal."Progressiveneuralnetworks."arXivpreprintarXiv:1606.04671(2016).[pdf](了不起的工作,新的想法)⭐️⭐️⭐️⭐️⭐️
2.8单样本学习
[59]Lake,BrenM.,RuslanSalakhutdinov,"Human-levelconceptlearningthroughprobabilisticprograminduction."(2015):1332-1338.[pdf](不是深度学习，但是值得一读)⭐️⭐️⭐️⭐️⭐️
[60]Koch,Gregory,RichardZemel,andRuslanSalakhutdinov."SiameseNeuralNetworksforOne-shotImageRecognition."(2015)[pdf]⭐️⭐️⭐️
[61]Santoro,Adam,etal."One-shotLearningwithMemory-AugmentedNeuralNetworks."arXivpreprintarXiv:1605.06065(2016).[pdf](一个单样本学习的基本步骤)⭐️⭐️⭐️⭐️
[62]Vinyals,Oriol,etal."MatchingNetworksforOneShotLearning."arXivpreprintarXiv:1606.04080(2016).[pdf]⭐️⭐️⭐️
[63]Hariharan,Bharath,andRossGirshick."Low-shotvisualobjectrecognition."arXivpreprintarXiv:1606.02819(2016).[pdf]⭐️⭐️⭐️⭐️
3应用3.1自然语言处理(NLP)
[1]AntoineBordes,etal."JointLearningofWordsandMeaningRepresentationsforOpen-TextSemanticParsing."AISTATS(2012)[pdf]⭐️⭐️⭐️⭐️
[2]Mikolov,etal."Distributedrepresentationsofwordsandphrasesandtheircompositionality."ANIPS(2013):3111-3119[pdf](词向量)⭐️⭐️⭐️
[3]Sutskever,etal."“Sequencetosequencelearningwithneuralnetworks."ANIPS(2014)[pdf]⭐️⭐️⭐️
[4]AnkitKumar,etal."“AskMeAnything:DynamicMemoryNetworksforNaturalLanguageProcessing."arXivpreprintarXiv:1506.07285(2015)[pdf]⭐️⭐️⭐️⭐️
[5]YoonKim,etal."Character-AwareNeuralLanguageModels."NIPS(2015)arXivpreprintarXiv:1508.06615(2015)[pdf]⭐️⭐️⭐️⭐️
[6]JasonWeston,etal."TowardsAI-CompleteQuestionAnswering:ASetofPrerequisiteToyTasks."arXivpreprintarXiv:1502.05698(2015)[pdf](问答任务)⭐️⭐️⭐️
[7]KarlMoritzHermann,etal."TeachingMachinestoReadandCompreh."arXivpreprintarXiv:1506.03340(2015)[pdf](CNN/DailyMailclozestylequestions)⭐️⭐️
[8]AlexisConneau,etal."VeryDeepConvolutionalNetworksforNaturalLanguageProcessing."arXivpreprintarXiv:1606.01781(2016)[pdf](文本分类的state-of-the-art的工作)⭐️⭐️⭐️
[9]ArmandJoulin,etal."BagofTricksforEfficientTextClassification."arXivpreprintarXiv:1607.01759(2016)[pdf](比state-of-the-art差点,但是快很多)⭐️⭐️⭐️
3.2物体检测
[1]Szegedy,Christian,AlexanderToshev,andDumitruErhan."Deepneuralnetworksforobjectdetection."[pdf]⭐️⭐️⭐️
[2]Girshick,Ross,etal."Richfeaturehierarchiesforaccurateobjectdetectionandsemanticsegmentation."Proc[pdf](RCNN)⭐️⭐️⭐️⭐️⭐️
[3]He,Kaiming,etal."Spatialpyramidpoolingindeepconvolutionalnetworksforvisualrecognition.",2014.[pdf](SPPNet)⭐️⭐️⭐️⭐️
[4]Girshick,Ross."Fastr-cnn."[pdf]⭐️⭐️⭐️⭐️
[5]Ren,Shaoqing,etal."FasterR-CNN:Towardsreal-timeobjectdetectionwithregionproposalnetworks."[pdf]⭐️⭐️⭐️⭐️
[6]Redmon,Joseph,etal."Youonlylookonce:Unified,real-timeobjectdetection."arXivpreprintarXiv:1506.02640(2015).[pdf](YOLO，实用)⭐️⭐️⭐️⭐️⭐️
[7]Liu,Wei,etal."SSD:SingleShotMultiBoxDetector."arXivpreprintarXiv:1512.02325(2015).[pdf]⭐️⭐️⭐️
[8]Dai,Jifeng,etal."R-FCN:ObjectDetectionviaRegion-basedFullyConvolutionalNetworks."arXivpreprintarXiv:1605.06409(2016).[pdf]⭐️⭐️⭐️⭐️
[9]He,Gkioxari,etal."MaskR-CNN"arXivpreprintarXiv:1703.06870(2017).[pdf]⭐️⭐️⭐️⭐️
3.3视觉跟踪
[1]Wang,Naiyan,andDit-YanYeung."Learningadeepcompactimagerepresentationforvisualtracking."[pdf](第一篇使用深度学习来进行视觉跟踪的论文，DLTTracker)⭐️⭐️⭐️
[2]Wang,Naiyan,etal."Transferringrichfeaturehierarchiesforrobustvisualtracking."arXivpreprintarXiv:1501.04587(2015).[pdf](SO-DLT)⭐️⭐️⭐️⭐️
[3]Wang,Lijun,etal."Visualtrackingwithfullyconvolutionalnetworks."[pdf](FCNT)⭐️⭐️⭐️⭐️
[4]Held,David,SebastianThrun,andSilvioSavarese."LearningtoTrackat100FPSwithDeepRegressionNetworks."arXivpreprintarXiv:1604.01802(2016).[pdf](GOTURN,作为深度学习的方法已经是很快的了，但是还是落后于非深度学习的方法)⭐️⭐️⭐️⭐️
[5]Bertinetto,Luca,etal."Fully-ConvolutionalSiameseNetworksforObjectTracking."arXivpreprintarXiv:1606.09549(2016).[pdf](SiameseFC,新的实时物体跟踪方面的state-of-the-art)⭐️⭐️⭐️⭐️
[6]MartinDanelljan,AndreasRobinson,FahadKhan,MichaelFelsberg."BeyondCorrelationFilters:LearningContinuousConvolutionOperatorsforVisualTracking."ECCV(2016)[pdf](C-COT)⭐️⭐️⭐️⭐️
[7]Nam,Hyeonseob,MooyeolBaek,andBohyungHan."ModelingandPropagatingCNNsinaTreeStructureforVisualTracking."arXivpreprintarXiv:1608.07242(2016).[pdf](VOT2016Winner,TCNN)⭐️⭐️⭐️⭐️
3.4图像描述
[1]Farhadi,Ali,etal."Everypicturetellsastory:Generatingsentencesfromimages".:15-29,2010.[pdf]⭐️⭐️⭐️
[2]Kulkarni,Girish,etal."Babytalk:Understandingandgeneratingimagedescriptions".InProceedingsofthe24thCVPR,2011.[pdf]⭐️⭐️⭐️⭐️
[3]Vinyals,Oriol,etal."Showandtell:Aneuralimagecaptiongenerator".InarXivpreprintarXiv:1411.4555,2014.[pdf]⭐️⭐️⭐️
[4]Donahue,Jeff,etal."Long-termrecurrentconvolutionalnetworksforvisualrecognitionanddescription".InarXivpreprintarXiv:1411.4389,2014.[pdf]
[5]Karpathy,Andrej,andLiFei-Fei."Deepvisual-semanticalignmentsforgeneratingimagedescriptions".InarXivpreprintarXiv:1412.2306,2014.[pdf]⭐️⭐️⭐️⭐️⭐️
[6]Karpathy,Andrej,ArmandJoulin,"Deepfragmentembeddingsforbidirectionalimagesentencemapping".InAdvancesinneuralinformationprocessingsystems,2014.[pdf]⭐️⭐️⭐️⭐️
[7]Fang,Hao,etal."Fromcaptionstovisualconceptsandback".InarXivpreprintarXiv:1411.4952,2014.[pdf]⭐️⭐️⭐️⭐️⭐️
[8]Chen,Xinlei,"Learningarecurrentvisualrepresentationforimagecaptiongeneration".InarXivpreprintarXiv:1411.5654,2014.[pdf]⭐️⭐️⭐️⭐️
[9]Mao,Junhua,etal."Deepcaptioningwithmultimodalrecurrentneuralnetworks(m-rnn)".InarXivpreprintarXiv:1412.6632,2014.[pdf]⭐️⭐️⭐️
[10]Xu,Kelvin,etal."Show,attandtell:Neuralimagecaptiongenerationwithvisualattention".InarXivpreprintarXiv:1502.03044,2015.[pdf]⭐️⭐️⭐️⭐️⭐️
3.5机器翻译
[1]Luong,Minh-Thang,etal."Addressingtherarewordprobleminneuralmachinetranslation."arXivpreprintarXiv:1410.8206(2014).[pdf]⭐️⭐️⭐️⭐️
[2]Sennrich,etal."NeuralMachineTranslationofRareWordswithSubwordUnits".InarXivpreprintarXiv:1508.07909,2015.[pdf]⭐️⭐️⭐️
[3]Luong,Minh-Thang,HieuPham,"Effectiveapproachestoattention-basedneuralmachinetranslation."arXivpreprintarXiv:1508.04025(2015).[pdf]⭐️⭐️⭐️⭐️
[4]Chung,etal."ACharacter-LevelDecoderwithoutExplicitSegmentationforNeuralMachineTranslation".InarXivpreprintarXiv:1603.06147,2016.[pdf]⭐️⭐️
[5]Lee,etal."FullyCharacter-LevelNeuralMachineTranslationwithoutExplicitSegmentation".InarXivpreprintarXiv:1610.03017,2016.[pdf]⭐️⭐️⭐️⭐️⭐️
[6]Wu,Schuster,Chen,Le,etal."Google'sNeuralMachineTranslationSystem:BridgingtheGapbetweenHumanandMachineTranslation".InarXivpreprintarXiv:1609.08144v2,2016.[pdf](里程碑)⭐️⭐️⭐️⭐️
3.6机器人
[1]Koutník,Jan,etal."Evolvinglarge-scaleneuralnetworksforvision-basedreinforcementlearning."Proceedi,2013.[pdf]⭐️⭐️⭐️
[2]Levine,Sergey,etal."-to-trainingofdeepvisuomotorpolicies."(2016):1-40.[pdf]⭐️⭐️⭐️⭐️⭐️
[3]Pinto,Lerrel,andAbhinavGupta."Supersizingself-supervision:Learningtograspfrom50ktriesand700robothours."arXivpreprintarXiv:1509.06825(2015).[pdf]⭐️⭐️⭐️
[4]Levine,Sergey,etal."LearningHand-EyeCoordinationforRoboticGraspingwithDeepLearningandLarge-ScaleDataCollection."arXivpreprintarXiv:1603.02199(2016).[pdf]⭐️⭐️⭐️⭐️
[5]Zhu,Yuke,etal."Target-drivenVisualNavigationinIndoorScenesusingDeepReinforcementLearning."arXivpreprintarXiv:1609.05143(2016).[pdf]⭐️⭐️⭐️⭐️
[6]Yahya,Ali,etal."CollectiveRobotReinforcementLearningwithDistributedAsynchronousGuidedPolicySearch."arXivpreprintarXiv:1610.00673(2016).[pdf]⭐️⭐️⭐️⭐️
[7]Gu,Shixiang,etal."DeepReinforcementLearningforRoboticManipulation."arXivpreprintarXiv:1610.00633(2016).[pdf]⭐️⭐️⭐️⭐️
[8]ARusu,MVecerik,ThomasRothörl,NHeess,RPascanu,RHadsell."Sim-to-RealRobotLearningfromPixelswithProgressiveNets."arXivpreprintarXiv:1610.04286(2016).[pdf]⭐️⭐️⭐️⭐️
[9]Mirowski,Piotr,etal."Learningtonavigateincomplexenvironments."arXivpreprintarXiv:1611.03673(2016).[pdf]⭐️⭐️⭐️⭐️
3.7艺术风格
[1]Mordvintsev,Alexander;Olah,Christopher;Tyka,Mike(2015)."Inceptionism:GoingDeeperintoNeuralNetworks".GoogleResearch.[html](DeepDream)⭐️⭐️⭐️⭐️
[2]Gatys,LeonA.,,andMatthiasBethge."Aneuralalgorithmofartisticstyle."arXivpreprintarXiv:1508.06576(2015).[pdf](杰出的工作,目前最成功的工作)⭐️⭐️⭐️⭐️⭐️
[3]Zhu,Jun-Yan,etal."GenerativeVisualManipulationontheNaturalImageManifold.",2016.[pdf](iGAN)⭐️⭐️⭐️⭐️
[4]Champandard,AlexJ."SemanticStyleTransferandTurningTwo-BitDoodlesintoFineArtworks."arXivpreprintarXiv:1603.01768(2016).[pdf](NeuralDoodle)⭐️⭐️⭐️⭐️
[5]Zhang,Richard,PhillipIsola,"ColorfulImageColorization."arXivpreprintarXiv:1603.08511(2016).[pdf]⭐️⭐️⭐️⭐️
[6]Johnson,Justin,AlexandreAlahi,andLiFei-Fei."Perceptuallossesforreal-timestyletransferandsuper-resolution."arXivpreprintarXiv:1603.08155(2016).[pdf]⭐️⭐️⭐️⭐️
[7]VincentDumoulin,JonathonShlensandManjunathKudlur."Alearnedrepresentationforartisticstyle."arXivpreprintarXiv:1610.07629(2016).[pdf]⭐️⭐️⭐️⭐️
[8]Gatys,LeonandEcker,etal."ControllingPerceptualFactorsinNeuralStyleTransfer."arXivpreprintarXiv:1611.07865(2016).[pdf](控制空间上的风格转换，颜色信息以及空间尺度)⭐️⭐️⭐️⭐️
[9]Ulyanov,DmitryandLebedev,Vadim,etal."TextureNetworks:Feed-forwardSynthesisofTexturesandStylizedImages."arXivpreprintarXiv:1603.03417(2016).[pdf](纹理生成和风格转换)⭐️⭐️⭐️⭐️
3.8物体分割
[1],,,“Fullyconvolutionalnetworksforsemanticsegmentation.”inCVPR,2015.[pdf]⭐️⭐️⭐️⭐️⭐️
[2]L.-,,,,"Semanticimagesegmentationwithdeepconvolutionalnetsandfullyconnectedcrfs."InICLR,2015.[pdf]⭐️⭐️⭐️⭐️⭐️
[3]Pinheiro,,Collobert,R.,Dollar,P."Learningtosegmentobjectcandidates."In:[pdf]⭐️⭐️⭐️⭐️
[4]Dai,J.,He,K.,Sun,J."Instance-awaresemanticsegmentationviamulti-tasknetworkcascades."[pdf]⭐️⭐️⭐️
[5]Dai,J.,He,K.,Sun,J."Instance-sensitiveFullyConvolutionalNetworks."arXivpreprintarXiv:1603.08678(2016).[pdf]⭐️⭐️⭐️
本文可以任意转载，转载时请注明作者及原文地址。
