Generating Cross Domain Models using Generative Adversarial Networks 
G. Victo Sudha George1, A. R. Raj Ganesh Arun2, S. Geetha3
1Dr. G. Victo Sudha George, Department of CSE, Dr. M.G.R. Educational and Research Institute, Chennai (Tamil Nadu), India.
2A.R. Raj Ganesh Arun, Department of CSE, Dr. M.G.R. Educational and Research Institute, Chennai (Tamil Nadu), India.
3Dr. S. Geetha, Department of CSE, Dr. M.G.R. Educational and Research Institute, Chennai (Tamil Nadu), India.
Manuscript received on 30 June 2019 | Revised Manuscript received on 05 July 2019 | Manuscript published on 30 July 2019 | PP: 1303-1306 | Volume-8 Issue-9, July 2019 | Retrieval Number: I8013078919/19©BEIESP | DOI: 10.35940/ijitee.I8013.078919
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Normally an average human can understand the relationship between different fields and domains easily with limited information exposure. On the other hand even a modern day computer is pretty bad in understanding relationship between different domains and fields which can be changed using a neural network called ‘GAN’ (Generative Adversarial Networks).GAN is a generative network that can generate data that we have ever seen before by Leveraging the generative nature of GAN . All of this is done using generator and discriminator in Generative Adversarial Networks. Two different datasets are trained on discriminator and generator. After training it generates new data which has attributes of both the datasets. But the biggest problem we face is with the training phase. Training in GAN is a like Min Max game. It doesn’t know when to stop. This is due to the fact that both Generator and discriminator try to overpower each other which leads to a lot of loss and error rate as it keeps on training. In this paper we have found an ideal state for the GAN.
Keywords: Neural Network, Deep Learning, GAN-Generative Adversarial Networks, GMM- Gaussian Mixture Model, Cross Domain Model
Scope of the Article: Neural Network