论文下载
bib:

@inproceedings{chenduan2016infogan,author       = {Xi Chen and Yan Duan and Rein Houthooft and John Schulman and Ilya Sutskever and Pieter Abbeel},title       = {InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets},booktitle     = {NIPS},year      = {2016},pages         = {2180--2188}
}

1. 摘要

This paper describes InfoGAN, an information-theoretic extension to the Generative Adversarial Network that is able to learn disentangled representations in a completely unsupervised manner. InfoGAN is a generative adversarial network that also maximizes the mutual information between a small subset of the latent variables and the observation. We derive a lower bound of the mutual information objective that can be optimized efficiently. Specifically, InfoGAN successfully disentangles writing styles from digit shapes on the MNIST dataset, pose from lighting of 3D rendered images, and background digits from the central digit on the SVHN dataset. It also discovers visual concepts that include hair styles, presence/absence of eyeglasses, and emotions on the CelebA face dataset. Experiments show that InfoGAN learns interpretable representations that are competitive with representations learned by existing supervised methods.

本文描述了InfoGAN，它是生成对抗网络的一个信息论扩展，能够以完全无监督的方式学习解耦表征。InfoGAN是一个生成式对抗网络，它也能最大化一个小子集的潜在变量与观测值之间的互信息。我们得到了一个可以有效优化的互信息目标的下界。具体来说，InfoGAN成功地从MNIST数据集的数字形状中分离出书写风格，从3D渲染图像的照明中分离出姿势，从SVHN数据集的中心数字中分离出背景数字。它还发现了CelebA脸部数据集上的视觉概念，包括发型、是否戴眼镜和情绪。实验表明，InfoGAN学习的可解释表示与现有的有监督学习方法学习的表示具有竞争性。

本文是一种GAN模型的变体，旨在学习一种可解释特征。

2. 前置知识

2.1 disentangled representations

Single latent units are sensitive to changes in single generative factors, while being relatively invariant to changes in other factors.

解耦表征学习是一个方向，意在获取一个表征，其中一个维度的变化对应于一个变化因子的变化，而其他因子相对不变。关于disentangled representations，可以参见博客。现在我的理解是，原本GAN对于一个潜在变量只能完成随机的生成任务，因为辨别器只能辨别生成器生成的图片是否为真实图片的二分类任务。在GAN的基础上，衍生出CGAN，即条件GAN，将标签带入生成。也就是辨别器不只是惩罚生成器不是真实图片的情况，还要惩罚生成图片不是对应标签的图片。按照现在的理解，InfoGAN更加进了一步，不只要控制生成图片的类别，还要控制生成图片的样式（style）。

2.2 mutual information

这里不涉及具体的数学解释，只是一个粗略的理解。互信息是一个随机变量由于已知另一个随机变量而减少的不肯定性。举个栗子，X=今天下雨，Y = 今天是阴天，那么已知道今天是阴天，那么X（今天下雨的概率）会增加，增加的量就是两个变量之间的互信息。由此可知，互信息只存在于两个非独立的随机变量中。Z=今天看论文，X与Z之间就没有互信息，每天都要看论文，与今天的天气无关

【论文阅读】InfoGAN: Interpretable Representation Learning by Information Maximizing GAN相关推荐

[生成对抗网络GAN入门指南]（10）InfoGAN: Interpretable Representation Learning by Information Maximizing GAN
本篇blog的内容基于原始论文InfoGAN: Interpretable Representation Learning by Information Maximizing Generative A ...
InfoGAN：Interpretable Representation Learning by Information Maximizing GANs论文解读
概述: InfoGAN是国际神经信息处理系统大会NIPS 2016上的论文,作者来自加州大学伯克利分校和OpenAI团队的研究人员,被OpenAI称为当年的五大突破之一.针对传统生成对抗网络以高度混杂 ...
InfoGAN：Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets ...
论文阅读笔记：SCAN: Learning to Classify Images without Labels
论文阅读笔记:SCAN: Learning to Classify Images without Labels 摘要简介和相关工作方法表征学习语义聚类损失 2.3 通过自标记进行微调 3 实验 ...
【论文阅读】Interpretable Visual Reasoning via Probabilistic Formulation under Natural Supervision
[论文阅读]Interpretable Visual Reasoning via Probabilistic Formulation under Natural Supervision 目录 [论文阅 ...
Zero-shot Learning零样本学习论文阅读（一）——Learning to detect unseen object classes by between-class attribute
Zero-shot Learning零样本学习论文阅读(一)--Learning to detect unseen object classes by between-class attribute ...
年龄论文阅读——Deep Label Distribution Learning With Label Ambiguity
论文阅读--Deep Label Distribution Learning With Label Ambiguity 版权声明:本文为博主原创文章,未经博主允许不得转载.https://blog.c ...
论文阅读|node2vec: Scalable Feature Learning for Networks
论文阅读|node2vec: Scalable Feature Learning for Networks 文章目录论文阅读|node2vec: Scalable Feature Learning ...
【论文阅读】Neural Transformation Learning for Deep Anomaly Detection Beyond Images 异常检测，可学习变换，时间序列，表格数据
本博客系博主阅读论文之后根据自己理解所写,非逐字逐句翻译,预知详情,请参阅论文原文. 论文标题:Neural Transformation Learning for Deep Anomaly Dete ...

【论文阅读】InfoGAN: Interpretable Representation Learning by Information Maximizing GAN

1. 摘要

2. 前置知识

2.1 disentangled representations

2.2 mutual information

【论文阅读】InfoGAN: Interpretable Representation Learning by Information Maximizing GAN相关推荐

最新文章

热门文章