| 李崇轩,朱军,张钹.条件概率图产生式对抗网络.软件学报,2020,31(4):0 |
| 条件概率图产生式对抗网络 |
| Conditional Graphical Generative Adverarial Networks |
| 投稿时间：2019-05-30 修订日期：2019-07-29 |
| DOI：10.13328/j.cnki.jos.005924 |
| 中文关键词: 深度生成模型 产生式对抗网络 概率图模型 弱监督学习 条件模型 |
| 英文关键词:deep generative models generative adversarial networks graphical models weakly-supervised leanring conditional models |
| 基金项目:国家自然科学基金（61620106010，61621136008） |
|
| 摘要点击次数: 64 |
| 全文下载次数: 31 |
| 中文摘要: |
| 产生式对抗网络（geneartive adverarial networks，GAN）可以生成逼真的图像，因此最近被广泛研究.值得注意的是，概率图生成对抗网络（Graphical-GAN）将贝叶斯网络引入产生式对抗网络框架，以无监督的方式学习到数据的隐藏结构.本文提出了条件概率图生成对抗网络（conditional Graphical-GAN），它可以在弱监督环境下，利用粗粒度监督信息来学习到更精细复杂的结构.条件概率图生成对抗网络的推理和学习遵循与Graphical-GAN类似的方法.我们提出了条件概率图生成对抗网络的两个实例.条件高斯混合模型（conditional Gaussian mixture GAN，cGMGAN）可以在给出粗粒度标签的情况下从混合数据中学习细粒度聚类.条件状态空间模型（conditional State Space GAN，cSSGAN）可以在给定对象标签的情况下学习具有多个对象的视频的动态过程. |
| 英文摘要: |
| Geneartive adverarial networks (GAN) have been promise on generating realistic images and hence have been studied widely. Notabaly, graphical generative adversarial networks (Graphical-GAN) introduce Bayesian networks to the GAN framework to learn the underlying structures of data in an unsupervised manner. In this paper, we propose a conditional version of Graphical-GAN, which can leverage coarse side information to enhance the Graphical-GAN and learn more fine and complex structures, in weakly-supervised learning settings. The inference and learning of conditional Graphical-GAN follows a similar protocol to Graphical-GAN. We present two instances of conditional Graphical-GAN. The conditional Gaussian mixture GAN can learn fine clusters from mixture data given a corse label. The conditional state space GAN can learn the dynamics of videos with multiple objects given the labels of the objects. |
|
HTML 下载PDF全文 查看/发表评论 下载PDF阅读器 |
