GAN model development
Project Details
- Project Lead
- Miao Jiang
- Project Manager
- Miao Jiang
- Project Members
- Dhawal Chaturvedi, B Fang
- Institution
- Indiana University, Intelligent Systems Engineering
- Discipline
- Computer Science (401)
- Subdiscipline
- 11.01 Computer and Information Sciences, General
Abstract
In this project we propose the Deli-Fisher GAN, a Generative Adversarial Network (GAN) that generates photo-realistic images by translating random data into meaningful data distributions.
In our framework, instead of using a single uniform or Gaussian distribution as source of random data,
we use mixture probability distributions to model the latent space and in addition, we include training on the distribution parameters. Furthermore, to improve stability and efficiency, we use the Fisher Integral Probability Metric as the divergence measure in our GAN model, instead of the usual Jensen Shannon divergence. We also show by experiments that our GAN performs no worse than the Fisher GAN, and that our GAN sometimes out-competes the Fisher GAN in terms of inception score while generating images with interesting properties.
Intellectual Merit
This project will have a impact on deep learning community. We have demonstrated our model from mathematics but need to make experiments. If the experiments are successful,we would try to publish the results into a conference.
Broader Impacts
Our proposed model could be applied to many applications in field of computer vision and artificial intelligence such as 3D reconstruction, inpainting and image recovering.
Scale of Use
We will use it from September 2018 to Dec,2018