arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1909.05063stat.MLcs.LG
收藏

基于独立子空间分析的非纠缠表示无监督学习

Independent Subspace Analysis for Unsupervised Learning of Disentangled Representations

Microsoft Research(微软研究院);University of Cambridge(剑桥大学);Google Brain(谷歌大脑)

Jan Stühmer, Richard E. Turner, Sebastian Nowozin

摘要:

最近,人们越来越关注使用变分自动编码器(VAE)框架对解纠缠表示进行无监督学习。大多数现有的工作主要集中在修改变分成本函数,以实现这一目标。我们首先表明,这些修改,例如β-VAE,简化的倾向,变分推断欠适应造成病理过度修剪和过度正交化的学习组件。其次,我们提出了一个补充的方法:修改概率模型与结构化的潜在先验。这种先验允许发现被构造成独立向量空间的层次结构的潜在变量表示。所提出的先验有三个主要优点:第一,与标准VAE正常先验相比,所提出的先验不是旋转不变的。这解决了标准VAE正态先验不可识别的问题。其次,我们证明了建议的先验鼓励解开的潜在表示,这有利于学习解开的表示。第三,大量的定量实验表明,在现有技术中,先验显著减轻了重建损失和解纠缠之间的权衡。

英文摘要:

Recently there has been an increased interest in unsupervised learning of disentangled representations using the Variational Autoencoder (VAE) framework. Most of the existing work has focused largely on modifying the variational cost function to achieve this goal. We first show that these modifications, e.g. beta-VAE, simplify the tendency of variational inference to underfit causing pathological over-pruning and over-orthogonalization of learned components. Second we propose a complementary approach: to modify the probabilistic model with a structured latent prior. This prior allows to discover latent variable representations that are structured into a hierarchy of independent vector spaces. The proposed prior has three major advantages: First, in contrast to the standard VAE normal prior the proposed prior is not rotationally invariant. This resolves the problem of unidentifiability of the standard VAE normal prior. Second, we demonstrate that the proposed prior encourages a disentangled latent representation which facilitates learning of disentangled representations. Third, extensive quantitative experiments demonstrate that the prior significantly mitigates the trade-off between reconstruction loss and disentanglement over the state of the art.

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。

↑