GANs are implicit generative networks. During a session at Quora, Yann LeCun, Director of AI Research at Facebook and Professor at NYU, described GANs as the most interesting idea in the last 10 years in ML. At present, lots of research is happening in GANs. Major AI/ML conferences conducted in the last few years have reported a majority of papers related to GANs.
GANs were proposed by Ian J. Goodfellow and Yoshua Bengio in the paper Generative Adversarial Networks in the year 2014 (https://arxiv.org/abs/1406.2661). They're inspired by the two-player game scenario. Like the two players of the game, in GANs, two networks—one called the discriminative network and the other the generative network—compete with each other. The generative network tries to generate data similar to the input data, and the discriminator network has to identify whether the data it's seeing is real or fake (that is, generated by a generator). Every time the discriminator finds a difference between the distribution...