對抗網(wǎng)絡(luò)專題文獻(xiàn)集

第一篇論文

[生成對抗網(wǎng)]（關(guān)于它的第一篇文章）

[紙張]：HTTPS：//arxiv.org/abs/1406.2661

[代碼]：HTTPS：//github.com/goodfeli/adversarial

未分類

[使用對角網(wǎng)絡(luò)的拉普拉斯金字塔的深度生成圖像模型]

[紙張] https://arxiv.org/abs/1506.05751

[代碼] https://github.com/facebook/eyescream

（具有深卷積生成對抗網(wǎng)絡(luò)的無監(jiān)督表示學(xué)習(xí)）（Gan與卷積網(wǎng)絡(luò)）（ICLR）

[紙張] https://arxiv.org/abs/1511.06434

[代碼] https://github.com/jacobgil/keras-dcgan

[對抗自動編碼器]

[紙張] http://arxiv.org/abs/1511.05644

[代碼] https://github.com/musyoku/adversarial-autoencoder

[基于深度網(wǎng)絡(luò)生成具有感知相似性度量的圖像]

[紙張] https://arxiv.org/pdf/1602.02644v2.pdf

[生成具有復(fù)發(fā)性對抗網(wǎng)絡(luò)的圖像]

[紙張] https://arxiv.org/abs/1602.05110

[代碼] https://github.com/ofirnachum/sequence_gan

[自然圖像歧管的生成視覺操作]

[紙張] https://people.eecs.berkeley.edu/%7Ejunyanz/projects/gvm/eccv16_gvm.pdf

[代碼] https://github.com/junyanz/iGAN

[生成對象文本到圖像合成]

[紙張] https://arxiv.org/abs/1605.05396

[代碼] https://github.com/reedscot/icml2016

[代碼] https://github.com/paarthneekhara/text-to-image

[學(xué)習(xí)什么和在哪里畫]

[紙張] http://www.scottreed.info/files/nips2016.pdf

[代碼] https://github.com/reedscot/nips2016

[草圖檢索對抗培訓(xùn)]

[紙張] http://link.springer.com/chapter/10.1007/978-3-319-46604-0_55

[使用風(fēng)格和結(jié)構(gòu)對抗網(wǎng)絡(luò)的生成圖像建模]

[紙張] https://arxiv.org/pdf/1603.05631.pdf

[代碼] https://github.com/xiaolonw/ss-gan

[生成對抗網(wǎng)絡(luò)作為能量模型的變化訓(xùn)練]（ICLR 2017）

[紙張] http://www.mathpubs.com/detail/1611.01799v1/Generative-Adversarial-Networks-as-Variational-Training-of-Energy-Based-Models

[半監(jiān)督文本分類對抗培訓(xùn)方法]（Ian Goodfellow Paper）

[紙張] https://arxiv.org/abs/1605.07725

[注意] https://github.com/dennybritz/deeplearning-papernotes/blob/master/notes/adversarial-text-classification.md

[通過對抗訓(xùn)練從模擬和無監(jiān)督的圖像學(xué)習(xí)]（蘋果論文）

[紙張] https://arxiv.org/abs/1612.07828

[代碼] https://github.com/carpedm20/simulated-unsupervised-tensorflow

[通過深層發(fā)電機(jī)網(wǎng)絡(luò)合成神經(jīng)網(wǎng)絡(luò)神經(jīng)元的首選輸入]

[紙張] https://arxiv.org/pdf/1605.09304v5.pdf

[代碼] https://github.com/Evolving-AI-Lab/synthesizing

[SalGAN：Visual Saliency Prediction with Generative Adversarial Networks]

[紙張] https://arxiv.org/abs/1701.01081

[代碼] https://github.com/imatge-upc/saliency-salgan-2017

[對抗特征學(xué)習(xí)]

[紙張] https://arxiv.org/abs/1605.09782

[使用循環(huán)一致性對抗網(wǎng)絡(luò)的無圖像到圖像轉(zhuǎn)換]

[紙張] https://junyanz.github.io/CycleGAN/

[代碼] https://github.com/junyanz/CycleGAN

合奏

[AdaGAN：Boosting Generative Models]（Google Brain）

[紙張] https://arxiv.org/abs/1701.02386

聚類

[使用生成對抗訓(xùn)練和聚類的無監(jiān)督學(xué)習(xí)]（ICLR）

[紙張] https://openreview.net/forum?id=SJ8BZTjeg?eId=SJ8BZTjeg

[代碼] https://github.com/VittalP/UnsupGAN

[無監(jiān)督和半監(jiān)督學(xué)習(xí)與分類生成對抗網(wǎng)絡(luò)]（ICLR）

[紙張] https://arxiv.org/abs/1511.06390

圖像修復(fù)

[感知和語境損失的語義圖像修復(fù)]

[紙張] https://arxiv.org/abs/1607.07539

[代碼] https://github.com/bamos/dcgan-completion.tensorflow

[上下文編碼器：通過修復(fù)進(jìn)行功能學(xué)習(xí)]

[紙張] https://arxiv.org/abs/1604.07379

[代碼] https://github.com/jazzsaxmafia/Inpainting

[上下文有條件生成對抗網(wǎng)絡(luò)的半監(jiān)督學(xué)習(xí)]

[紙張] https://arxiv.org/abs/1611.06430v1

聯(lián)合概率

[對峙學(xué)習(xí)推論]

[紙張] https://arxiv.org/abs/1606.00704

[代碼] https://github.com/IshmaelBelghazi/ALI

超分辨率

[通過深度學(xué)習(xí)的圖像超分辨率]（僅面向數(shù)據(jù)集）

[代碼] https://github.com/david-gpu/srez

[使用生成對抗網(wǎng)絡(luò)的照片逼真單圖像超分辨率]（使用深度殘差網(wǎng)絡(luò)）

[紙張] https://arxiv.org/abs/1609.04802

[代碼] https://github.com/leehomyc/Photo-Realistic-Super-Resoluton

[EnhanceGAN]

[文件] https://medium.com/@richardherbert/faces-from-noise-super-enhancing-8x8-images-with-enhancegan-ebda015bb5e0#.io6pskvin

去除遮蔽

[強(qiáng)大的LSTM自動編碼器在野外面部遮擋]

[紙張] https://arxiv.org/abs/1612.08534

語義分割

[使用對話網(wǎng)絡(luò)的語義分割]（soumith的論文）

[紙張] https://arxiv.org/abs/1611.08408

對象檢測

[用于小物體檢測的感知生成對抗網(wǎng)絡(luò)]（提交）

[A-Fast-RCNN：通過對象檢測的對手的硬正產(chǎn)生]（CVPR2017）

[紙] http://abhinavsh.info/papers/pdfs/adversarial_object_detection.pdfRNN

[C-RNN-GAN：具有對抗性訓(xùn)練的連續(xù)循環(huán)神經(jīng)網(wǎng)絡(luò)]

[紙張] https://arxiv.org/abs/1611.09904

[代碼] https://github.com/olofmogren/c-rnn-gan

有條件的對抗

[有條件生成對抗網(wǎng)]

[紙張] https://arxiv.org/abs/1411.1784

[代碼] https://github.com/zhangqianhui/Conditional-Gans

[InfoGAN：Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets]

[紙張] https://arxiv.org/abs/1606.03657

[代碼] https://github.com/buriburisuri/supervised_infogan

[使用條件對抗網(wǎng)絡(luò)的圖像到圖像翻譯]

[紙張] https://arxiv.org/pdf/1611.07004v1.pdf

[代碼] https://github.com/phillipi/pix2pix

[代碼] https://github.com/yenchenlin/pix2pix-tensorflow

[使用輔助分類器GAN的條件圖像合成]（GoogleBrain ICLR 2017）

[紙張] https://arxiv.org/abs/1610.09585

[代碼] https://github.com/buriburisuri/ac-gan

[像素級域名轉(zhuǎn)移]

[紙張] https://arxiv.org/pdf/1603.07442v2.pdf

[代碼] https://github.com/fxia22/pldtgan

[圖像編輯的可逆條件GAN]

[紙張] https://arxiv.org/abs/1611.06355

[代碼] https://github.com/Guim3/IcGAN

[即插即用生成網(wǎng)絡(luò)：潛在空間中的條件迭代生成圖像]

[紙張] https://arxiv.org/abs/1612.00005v1

[代碼] https://github.com/Evolving-AI-Lab/ppgn

[StackGAN：文本到具有堆疊生成對話網(wǎng)絡(luò)的照片逼真圖像合成]

[紙張] https://arxiv.org/pdf/1612.03242v1.pdf

[代碼] https://github.com/hanzhanggit/StackGAN

[無監(jiān)督的圖像到圖像翻譯與生成對抗網(wǎng)絡(luò)]

[紙張] https://arxiv.org/pdf/1701.02676.pdf

[學(xué)習(xí)與生成對話網(wǎng)絡(luò)發(fā)現(xiàn)跨域關(guān)系]

[紙張] https://arxiv.org/abs/1703.05192

[代碼] https://github.com/carpedm20/DiscoGAN-pytorch

視頻預(yù)測

[深度多尺度視頻預(yù)測超過均方誤差]（Yann LeCun的論文）

[紙張] https://arxiv.org/abs/1511.05440

[代碼] https://github.com/dyelax/Adversarial_Video_Generation

[通過視頻預(yù)測進(jìn)行物理互動的無監(jiān)督學(xué)習(xí)]（Ian Goodfellow的論文）

[紙張] https://arxiv.org/abs/1605.07157

[使用場景動態(tài)生成視頻]

[紙張] https://arxiv.org/abs/1609.02612

[網(wǎng)絡(luò)] http://web.mit.edu/vondrick/tinyvideo/

[代碼] https://github.com/cvondrick/videogan

紋理合成和風(fēng)格轉(zhuǎn)移

[使用馬爾可夫生成對抗網(wǎng)絡(luò)的預(yù)計(jì)算實(shí)時(shí)紋理合成]（ECCV 2016）

[紙張] https://arxiv.org/abs/1604.04382

[代碼] https://github.com/chuanli11/MGANs

GAN理論

[能源生成對抗網(wǎng)]（Lecun論文）

[紙張] https://arxiv.org/pdf/1609.03126v2.pdf

[代碼] https://github.com/buriburisuri/ebgan

[改進(jìn)GAN培訓(xùn)技巧]（Goodfellow的論文）

[紙張] https://arxiv.org/abs/1606.03498

[代碼] https://github.com/openai/improved-gan

[模式正則化生成對抗網(wǎng)絡(luò)]（Yoshua Bengio，ICLR 2017）

[紙張] https://openreview.net/pdf?id=HJKkY35le

[改進(jìn)產(chǎn)生對抗網(wǎng)絡(luò)的去噪特征匹配]（Yoshua Bengio立砸，ICLR 2017）

[紙張] https://openreview.net/pdf?id=S1X7nhsxl

[代碼] https://github.com/hvy/chainer-gan-denoising-feature-matching

[采樣生成網(wǎng)絡(luò)]

[紙張] https://arxiv.org/abs/1609.04468

[代碼] https://github.com/dribnet/plat

[模式正則化生成對話網(wǎng)絡(luò)]（Yoshua Bengio的論文）

[紙張] https://arxiv.org/abs/1612.02136

[如何訓(xùn)練甘斯]

[的Docu] https://github.com/soumith/ganhacks#authors

[面向訓(xùn)練生成對抗網(wǎng)絡(luò)的原則方法]（ICLR 2017）

[紙張] http://openreview.net/forum?id=Hk4_qw5xe

[展開的生成對抗網(wǎng)絡(luò)]

[紙張] https://arxiv.org/abs/1611.02163

[代碼] https://github.com/poolio/unrolled_gan

[最小二乘法對抗網(wǎng)絡(luò)]

[紙張] https://arxiv.org/abs/1611.04076

[代碼] https://github.com/pfnet-research/chainer-LSGAN

[Wasserstein GAN]

[紙張] https://arxiv.org/abs/1701.07875

[代碼] https://github.com/martinarjovsky/WassersteinGAN

[Lipschitz密度損失敏感的生成對抗網(wǎng)絡(luò)]（與WGan相同）

[紙張] https://arxiv.org/abs/1701.06264

[代碼] https://github.com/guojunq/lsgan

[面向訓(xùn)練生成對抗網(wǎng)絡(luò)的原則方法]

[紙張] https://arxiv.org/abs/1701.04862

3D

[通過3D生成 - 對抗建模學(xué)習(xí)對象形狀的概率潛在空間]（2016 NIPS）

[紙張] https://arxiv.org/abs/1610.07584

[網(wǎng)絡(luò)] http://3dgan.csail.mit.edu/

[代碼] https://github.com/zck119/3dgan-release

面對生成和編輯

[使用學(xué)習(xí)的相似性度量自動編碼超像素

[紙張] https://arxiv.org/abs/1512.09300

[代碼] https://github.com/andersbll/autoencoding_beyond_pixels

[耦合生成對抗網(wǎng)絡(luò)]（NIPS）

[紙張] http://mingyuliu.net/

[Caffe Code] https://github.com/mingyuliutw/CoGAN

[Tensorflow Code] https://github.com/andrewliao11/CoGAN-tensorflow

[圖像編輯的可逆條件GAN]

[紙張] https://drive.google.com/file/d/0B48XS5sLi1OlRkRIbkZWUmdoQmM/view

[代碼] https://github.com/Guim3/IcGAN

[面部屬性操縱的學(xué)習(xí)殘差圖像]

[紙張] https://arxiv.org/abs/1612.05363

[使用內(nèi)省對抗網(wǎng)絡(luò)進(jìn)行神經(jīng)照片編輯]（ICLR 2017）

[紙張] https://arxiv.org/abs/1609.07093

[代碼] https://github.com/ajbrock/Neural-Photo-Editor

對于離散分布

[最大似然增強(qiáng)離散生成對抗網(wǎng)絡(luò)]

[紙張] https://arxiv.org/abs/1702.07983v1

[邊界尋求生成對抗網(wǎng)絡(luò)]

[紙張] https://arxiv.org/abs/1702.08431

[GANS-GANSB]的分離元素序列與Gumbel-softmax分布

[紙張] https://arxiv.org/abs/1611.04051

項(xiàng)目

[cleverhans]（一個(gè)用于對抗脆弱性的對抗圖書館）

[代碼] https://github.com/openai/cleverhans

[reset-cppn-gan-tensorflow]（使用殘差生成對抗網(wǎng)絡(luò)和變分自動編碼器技術(shù)來產(chǎn)生高分辨率圖像）

[代碼] https://github.com/hardmaru/resnet-cppn-gan-tensorflow

（HyperGAN）（開源GAN著重于規(guī)模和可用性）

[代碼] https://github.com/255bits/HyperGAN