新智元启动 2017 最新一轮大招聘:。 新智元为COO和执行总编提供最高超百万的年薪激励;为骨干员工提供最完整的培训体系、高于业界平均水平的工资和奖金。加盟新智元,与人工智能业界领袖携手改变世界。 简历投递:j[email protected] HR 微信:13552313024 【新智元导读】这是一份生成对抗(神经)网络的重要论文以及其他资源的列表,由 Holger Caesar 整理,包括重要的 workshops,教程和博客,按主题分类的重要论文,视频,代码等,值得收藏学习。 目录 Workshops 教程 & 博客 论文 理论 & 机器学习 视觉应用 其他应用 幽默 Workshops NIP 2016 对抗训练 Workshop 【网页】https://sites.google.com/site/nips2016adversarial/ 【博客】 教程 & 博客 如何训练 GAN? 让 GAN 工作的提示和技巧 【博客】https://github.com/soumith/ganhacks NIPS 2016 教程:生成对抗网络 【arXiv】https://arxiv.org/abs/1701.00160 深度学习和 GAN 背后的直觉知识——一个基础理解 【博客】https://blog.waya.ai/introduction-to-gans-a-boxing-match-b-w-neural-nets-b4e5319cc935 OpenAI——生成模型 【博客】https://openai.com/blog/generative-models/ SimGANs——无监督学习的游戏规则颠覆者,无人车等 【博客】https://blog.waya.ai/simgans-applied-to-autonomous-driving-5a8c6676e36b 论文 理论 & 机器学习 生成对抗网络,逆向强化学习和 Energy-Based 模型之间的联系(A Connection between Generative Adversarial Networks, Inverse Reinforcement Learning, and Energy-Based Models ) 可扩展对抗分类的通用训练框架(A General Retraining Framework for Scalable Adversarial Classification) 对抗自编码器(Adversarial Autoencoders) 对抗判别的领域适应(Adversarial Discriminative Domain Adaptation) 对抗性 Generator-Encoder 网络(Adversarial Generator-Encoder Networks) 对抗特征学习(Adversarial Feature Learning) 【代码】https://github.com/wiseodd/generative-models 对抗推理学习(Adversarially Learned Inference) 【代码】https://github.com/wiseodd/generative-models 结构化输出神经网络半监督训练的一种对抗正则化(An Adversarial Regularisation for Semi-Supervised Training of Structured Output Neural Networks) 联想式对抗网络(Associative Adversarial Networks) b-GAN:生成对抗网络的一个新框架(b-GAN: New Framework of Generative Adversarial Networks) 【代码】https://github.com/wiseodd/generative-models 边界寻找生成对抗网络(Boundary-Seeking Generative Adversarial Networks) 【代码】https://github.com/wiseodd/generative-models 条件生成对抗网络(Conditional Generative Adversarial Nets) 【代码】https://github.com/wiseodd/generative-models 结合生成对抗网络和 Actor-Critic 方法(Connecting Generative Adversarial Networks and Actor-Critic Methods) 描述符和生成网络的协同训练(Cooperative Training of Deor and Generator Networks) Coupled Generative Adversarial Networks(CoGAN) 【代码】https://github.com/wiseodd/generative-models 基于能量模型的生成对抗网络(Energy-based Generative Adversarial Network) 【代码】https://github.com/wiseodd/generative-models 对抗样本的解释和利用(Explaining and Harnessing Adversarial Examples) f-GAN:使用变分发散最小化训练生成式神经采样器(f-GAN: Training Generative Neural Samplers using Variational Divergence Minimization) Gang of GANs: Generative Adversarial Networks with Maximum Margin Ranking 用递归对抗网络乘车图像(Generating images with recurrent adversarial networks) Generative Adversarial Nets with Labeled Data by Activation Maximization 生成对抗网络(Generative Adversarial Networks) 【代码】https://github.com/goodfeli/adversarial 【代码】https://github.com/wiseodd/generative-models 生成对抗并行化(Generative Adversarial Parallelization) 【代码】https://github.com/wiseodd/generative-models One Shot学习的生成性对抗残差成对网络(Generative Adversarial Residual Pairwise Networks for One Shot Learning) 生成对抗结构化网络(Generative Adversarial Structured Networks) 生成式矩匹配网络(Generative Moment Matching Networks) 【代码】https://github.com/yujiali/gmmn 训练GAN的改进技术(Improved Techniques for Training GANs) 【代码】https://github.com/openai/improved-gan 改善训练WGAN(Improved Training of Wasserstein GANs) 【代码】https://github.com/wiseodd/generative-models InfoGAN:通过信息最大化GAN学习可解释表示(InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets) 【代码】https://github.com/wiseodd/generative-models 翻转GAN的生成器(Inverting The Generator Of A Generative Adversarial Network) 隐式生成模型里的学习(Learning in Implicit Generative Models) 用GAN学习发现跨域关系(Learning to Discover Cross-Domain Relations with Generative Adversarial Networks) 【代码】https://github.com/wiseodd/generative-models 最小二乘生成对抗网络,LSGAN(Least Squares Generative Adversarial Networks) 【代码】https://github.com/wiseodd/generative-models LS-GAN,损失敏感GAN(Loss-Sensitive Generative Adversarial Networks on Lipschitz Densities) LR-GAN:用于图像生成的分层递归GAN(LR-GAN: Layered Recursive Generative Adversarial Networks for Image Generation) MAGAN: Margin Adaptation for Generative Adversarial Networks 【代码】https://github.com/wiseodd/generative-models 最大似然增强的离散生成对抗网络(Maximum-Likelihood Augmented Discrete Generative Adversarial Networks) 模式正则化GAN(Mode Regularized Generative Adversarial Networks) 【代码】https://github.com/wiseodd/generative-models Multi-Agent Diverse Generative Adversarial Networks 生成对抗网络中Batch Normalization和Weight Normalization的影响(On the effect of Batch Normalization and Weight Normalization in Generative Adversarial Networks) 基于解码器的生成模型的定量分析(On the Quantitative Analysis of Decoder-Based Generative Models) SeqGAN:策略渐变的序列生成对抗网络(SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient) 深度网络的简单黑箱对抗干扰(Simple Black-Box Adversarial Perturbations for Deep Networks) Stacked GAN(Stacked Generative Adversarial Networks) 通过最大均值差异优化训练生成神经网络(Training generative neural networks via Maximum Mean Discrepancy optimization) Triple Generative Adversarial Nets Unrolled Generative Adversarial Networks DCGAN无监督表示学习(Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks) 【代码】https://github.com/Newmu/dcgan_code 【代码】https://github.com/pytorch/examples/tree/master/dcgan 【代码】https://github.com/carpedm20/DCGAN-tensorflow 【代码】https://github.com/jacobgil/keras-dcgan Wasserstein GAN(WGAN) 【代码】https://github.com/martinarjovsky/WassersteinGAN 【代码】https://github.com/wiseodd/generative-models 视觉应用 用对抗网络检测恶性前列腺癌(Adversarial Networks for the Detection of Aggressive Prostate Cancer) 条件对抗自编码器的年龄递进/回归(Age Progression / Regression by Conditional Adversarial Autoencoder) ArtGAN:条件分类GAN的艺术作品合成(ArtGAN: Artwork Synthesis with Conditional Categorial GANs) Beyond Face Rotation: Global and Local Perception GAN for Photorealistic and Identity Preserving Frontal View Synthesis 卷积人脸生成的条件GAN(Conditional generative adversarial nets for convolutional face generation) 辅助分类器GAN的条件图像合成(Conditional Image Synthesis with Auxiliary Classifier GANs) 【代码】https://github.com/wiseodd/generative-models 使用对抗网络的Laplacian金字塔的深度生成图像模型(Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks) 【代码】https://github.com/facebook/eyescream 【博客】 Deep multi-scale video prediction beyond mean square error 【代码】https://github.com/dyelax/Adversarial_Video_Generation DualGAN:图像到图像翻译的无监督Dual学习(DualGAN: Unsupervised Dual Learning for Image-to-Image Translation) 【代码】https://github.com/wiseodd/generative-models 用循环神经网络做全分辨率图像压缩(Full Resolution Image Compression with Recurrent Neural Networks) 生成以适应:使用GAN对齐域(Generate To Adapt: Aligning Domains using Generative Adversarial Networks) 生成对抗文本到图像的合成(Generative Adversarial Text to Image Synthesis) 【代码】https://github.com/paarthneekhara/text-to-image 自然图像流形上的生成视觉操作(Generative Visual Manipulation on the Natural Image Manifold) 【项目】~junyanz/projects/gvm/ 【视频】https://youtu.be/9c4z6YsBGQ0 【代码】https://github.com/junyanz/iGAN Image De-raining Using a Conditional Generative Adversarial Network Image Generation and Editing with Variational Info Generative Adversarial Networks 用条件对抗网络做 Image-to-Image 翻译(Image-to-Image Translation with Conditional Adversarial Networks) 【代码】https://github.com/phillipi/pix2pix 用GAN模仿驾驶员行为(Imitating Driver Behavior with Generative Adversarial Networks) 可逆的条件GAN用于图像编辑(Invertible Conditional GANs for image editing) 学习驱动模拟器(Learning a Driving Simulator) 多视角GAN(Multi-view Generative Adversarial Networks) 利用内省对抗网络编辑图片(Neural Photo Editing with Introspective Adversarial Networks) 使用GAN生成照片级真实感的单一图像超分辨率(Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network) Recurrent Topic-Transition GAN for Visual Paragraph Generation RenderGAN:生成现实的标签数据(RenderGAN: Generating Realistic Labeled Data) SeGAN: Segmenting and Generating the Invisible 使用对抗网络做语义分割(Semantic Segmentation using Adversarial Networks) 半隐性GAN:学习从特征生成和修改人脸图像(Semi-Latent GAN: Learning to generate and modify facial images from attributes) TAC-GAN - 文本条件辅助分类器GAN(TAC-GAN - Text Conditioned Auxiliary Classifier Generative Adversarial Network) 通过条件GAN实现多样化且自然的图像描述(Towards Diverse and Natural Image Deions via a Conditional GAN) GAN 提高人的体外识别基线的未标记样本生成(Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro) Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks 无监督异常检测,用GAN指导标记发现(Unsupervised Anomaly Detection with Generative Adversarial Networks to Guide Marker Discovery) 无监督跨领域图像生成(Unsupervised Cross-Domain Image Generation) (责任编辑:本港台直播) |