Back to listing

Generative Adversarial Networks (GANs)

Author Picture

Ayush Kanchan


GANs as we call them in the technology world are a form of neural network model that function on the principle of unsupervised machine learning. This means that the network model automatically discovers and learns the regular patterns in the input data in such a way that it can generate new results that could have been drawn from the original dataset that has been used. The interpretation is made in a way that the same model can be used to process new examples.

For instance: GANs can create images/ pictures that look like the photographs of human faces, even though the faces don't belong to any real person.

The underlying idea of the functioning of GANs is to redesign virtual levels and worlds by cloning the operations of a human being.

The GAN model involves two sub-models: a Generator model for generating new instances and a Discriminator model for analyzing whether the generated examples are real, from the domain, or reproduced by the generator model. The application of GANs has increased rapidly over the years including News forecasting, Agriculture sector, Fashion industry, Food Industry, Science, Advertisement sports industry, etc.