arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1906.09333cs.LGcs.AI
收藏

SeGMA:半监督高斯混合自动编码器

SeGMA: Semi-Supervised Gaussian Mixture Auto-Encoder

Jagiellonian University(雅盖隆大学)

Marek Śmieja, Maciej Wołczyk, Jacek Tabor, Bernhard C. Geiger

摘要:

我们提出了一个半监督生成模型,SeGMA,它学习数据和它们的类的联合概率分布,并在一个典型的Wasserstein自动编码器框架中实现。我们选择高斯的混合作为潜在空间中的目标分布,这提供了将数据自然分割成聚类的方法。为了将高斯分量与正确的类连接起来,我们使用少量的标记数据和由目标分布引起的高斯分类器。SeGMA是有效地优化,由于使用Cramer-Wold距离作为最大平均差异惩罚,这产生了一个封闭形式的表达式的混合球高斯分量,从而消除了采样的需要。虽然SeGMA保留了其半监督前辈的所有属性,并在标准基准数据集上实现了至少同样好的生成性能,但它提供了额外的功能:(a)潜在空间中任何一对点之间的插值产生逼真的样本;(b)将插值属性与分离的类和样式变量相结合,SeGMA能够执行从一个类到另一个类的连续风格转移;(c)通过将数据点的潜在表示从特定的高斯分量移开,可以改变数据点中类特征的强度。

英文摘要:

We propose a semi-supervised generative model, SeGMA, which learns a joint probability distribution of data and their classes and which is implemented in a typical Wasserstein auto-encoder framework. We choose a mixture of Gaussians as a target distribution in latent space, which provides a natural splitting of data into clusters. To connect Gaussian components with correct classes, we use a small amount of labeled data and a Gaussian classifier induced by the target distribution. SeGMA is optimized efficiently due to the use of Cramer-Wold distance as a maximum mean discrepancy penalty, which yields a closed-form expression for a mixture of spherical Gaussian components and thus obviates the need of sampling. While SeGMA preserves all properties of its semi-supervised predecessors and achieves at least as good generative performance on standard benchmark data sets, it presents additional features: (a) interpolation between any pair of points in the latent space produces realistically-looking samples; (b) combining the interpolation property with disentangled class and style variables, SeGMA is able to perform a continuous style transfer from one class to another; (c) it is possible to change the intensity of class characteristics in a data point by moving the latent representation of the data point away from specific Gaussian components.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑