DCGAN-论文阅读笔记
论文结构
1.Introduction
2.Related work
2.1 Representation learning from unlabeled data
2.2 Generating natural images
2.3 Visualizing the internals of CNNs
3.Approach and model architecture
4.Details of adversarial training
4.1 LSUN
4.1.1 Deduplication
4.2 Faces
4.3 Imagenet-1k
5.Empirical validation of DCGANs capabilities
5.1 Classifying CIFAR-10 using GANs as a feature extractor
5.2 Classifying SVHN digits using GANs as a feature
extractor
6.Investigating and visualizing the internals of the networks
6.1 Walking in the latent space
6.2 Visualizing the discriminator features
6.3 Manipulating the generator representation
6.3.1 Forgetting to draw certain objects
6.3.2 Vector arithmetic on face samples
7.Conclusion and future work
摘要
Abstract: In recent years, supervised learning with convolutional networks (CNNs) has seen huge adoption in computer vision applications. Comparatively, unsupervised learning with CNNs has received less attention. In this work we hope to help bridge the gap between the success of CNNs for supervised learning and unsupervised learning. We introduce a class of CNNs called deep convolutional generative adversarial networks (DCGANs), that have certain architectural constraints, and demonstrate that they are a strong candidate for unsupervised learning. Training on various image datasets, we show convincing evidence that our deep convolutional adversarial pair learns a hierarchy of representations from object parts to scenes in both the generator and discriminator. Additionally, we use the learned features for novel tasks - demonstrating their applicability as general image representations
摘要核心
1.希望能让CNN在无监督学习上，达到与监督学习一样的成功
2.通过架构约束，构建了深度卷积生成对抗网络（DCGAN）
3.证明了DCGAN是目前先进的无监督学习网络
4.证明了DCGAN的生成器和判别器学习到了从物体细节到整体场景的多层次表征
5.证明了DCGAN判别器提取的图像特征具有很好的泛化性
研究背景
表征学习
•表征（representation）、特征（feature）、编码（code）
好的表征
具有很强的表示能力，即同样大小的向量可以表示更多信息
使后续的学习任务变得简单，即需要包含更高层的语义信息
具有泛化性，可以应用到不同领域
表征学习的方式
无监督表征学习
有监督表征学习
模型可解释性
—— Interpretation is the process of giving explanations to Human
•
决策树就是一个具有良好可解释性的模型
•
使用特征可视化方法
•
使用数据分析，可以找到数据中一些具有代表性和不具代表性的样本
•
NIPS 2017会议上，Yann LeCun：人类大脑是非常有限的，我们没有那么多脑容量去研究所有东西的可解释性
研究成果
