Timezone: »

 
Poster
Structured Generative Adversarial Networks
Zhijie Deng · Hao Zhang · Xiaodan Liang · Luona Yang · Shizhen Xu · Jun Zhu · Eric Xing

Wed Dec 06 06:30 PM -- 10:30 PM (PST) @ Pacific Ballroom #117
in Posters Wed »

We study the problem of conditional generative modeling based on designated semantics or structures. Existing models that build conditional generators either require massive labeled instances as supervision or are unable to accurately control the semantics of generated samples. We propose structured generative adversarial networks (SGANs) for semi-supervised conditional generative modeling. SGAN assumes the data x is generated conditioned on two independent latent variables: y that encodes the designated semantics, and z that contains other factors of variation. To ensure disentangled semantics in y and z, SGAN builds two collaborative games in the hidden space to minimize the reconstruction error of y and z, respectively. Training SGAN also involves solving two adversarial games that have their equilibrium concentrating at the true joint data distributions p(x, z) and p(x, y), avoiding distributing the probability mass diffusely over data space that MLE-based methods may suffer. We assess SGAN by evaluating its trained networks, and its performance on downstream tasks. We show that SGAN delivers a highly controllable generator, and disentangled representations; it also establishes start-of-the-art results across multiple datasets when applied for semi-supervised image classification (1.27%, 5.73%, 17.26% error rates on MNIST, SVHN and CIFAR-10 using 50, 1000 and 4000 labels, respectively). Benefiting from the separate modeling of y and z, SGAN can generate images with high visual quality and strictly following the designated semantic, and can be extended to a wide spectrum of applications, such as style transfer.

Keywords: Semi-Supervised Learning · Deep Autoencoders · Adversarial Networks · Generative Models
[ Code]  [ Paper]  [ Poster] 

Author Information

Zhijie Deng (Tsinghua University)
Hao Zhang (Carnegie Mellon University)
Xiaodan Liang (Sun Yat-sen University)
Luona Yang (Carnegie Mellon University)
Shizhen Xu (Tsinghua University)
Jun Zhu (Tsinghua University)
Eric Xing (Petuum Inc. / Carnegie Mellon University)

More from the Same Authors