[Viewpoint] In the various learning of artificial intelligence, perhaps the adversarial learning of adversarial networks is the next hot spot after deep learning and enhanced learning. It subverts the paradigm of deep learning to a certain extent, disrupting the division of categories and their categories before machine learning. Focusing on the frontiers of artificial intelligence, it is necessary to understand the generation of adversarial networks (defined below as adversarial learning)-as described in "A Brief History of the New Future".
In machine learning algorithms including deep learning, reinforcement learning, adversarial learning, and transfer learning (which are basically impossible today), the adversarial learning method of "generative adversarial networks" is very important. This pioneering new learning idea is likely to Become the next hot spot in deep learning. It is necessary to discuss this in depth here.
Subverting the deep learning paradigm
Ever since Ian Goodfellow published the paper "Generative Adversarial Nets" (GANs) in 2014, adversarial learning has attracted much attention. In addition, the global academic bull, Facebook Artificial Intelligence Yann LeCun, the head of the research center FAIR and one of the deep learning troikas, said when he answered the question online:
"His most exciting progress in deep learning is generative adversarial networks," making adversarial learning (GAN) a new favorite in the field of machine learning in recent years.
On June 18, 2017, Jan LeChun retweeted a post on Facebook, introducing a research result he was one of the collaborators: