Skip to content

Generative Neural Nets

Generative Adversarial Networks (or GANs) have gained lot of popularity in recent years to generate new artificial data by feeding the deep learning model with real data samples.. The model learns to create new samples by following the same distribution or patterns as in the real data-set used for training..

Applications of GANs in Media & Entertainment domain include:

  • Character Generation: To generate new characters for video games & animation films, by feeding the model with a sample of human-drawn characters like cartoons, caricatures, comic book action figures etc..
  • Automatic Scene Generation: To create new background scenery by training the machine learning models on sample images and pictures of nature & landscape photography
  • Music Generation: To generate new musical patterns or scores similar to the given training sample