{"id":8065,"date":"2018-02-09T07:00:26","date_gmt":"2018-02-09T15:00:26","guid":{"rendered":"http:\/\/formtek.com\/blog\/?p=8065"},"modified":"2017-12-22T07:43:36","modified_gmt":"2017-12-22T15:43:36","slug":"artificial-intelligence-autonomous-unsupervised-machine-intelligence","status":"publish","type":"post","link":"https:\/\/formtek.com\/blog\/artificial-intelligence-autonomous-unsupervised-machine-intelligence\/","title":{"rendered":"Artificial Intelligence: Autonomous Unsupervised Machine Intelligence"},"content":{"rendered":"

“Generative Adversarial Networks” (GAN)<\/a> are two AI programs that work side-by-side, the first is designed to generate something like an image or music, and the second algorithm judges the results of the first program. The feedback from the judging algorithm helps the first algorithm improve while the second program also improves in its ability to recognize better quality results.\u00a0 Over time, both programs improve.\u00a0 GANS are an attempt to create unsupervised AI — algorithms that can work autonomously without input from humans.<\/p>\n

Jay Garmon<\/a>, tech writer, told TechRepublic that<\/a> “you tend to have two different AI systems in this pit fight: One is the generator and one is the discriminator. The better the generator gets at fooling the discriminator, you can take that data and retrain the discriminator to get better at spotting fakes. Back and forth it goes until both are really, really good at their jobs, without you having to go out and get 10,000 pictures of stop signs.”<\/p>\n

Some applications of GANs<\/a> are not that surprising and include:<\/p>\n