3 ms·
Gen Studio – An experimental collaboration across The Met, Microsoft, and MIT
- najarvg 8y agoThis is a very interesting visualization of GAN capabilities. This seems a part of a larger group of collaborative projects between the MET, MIT and MS - https://www.microsoft.com/inculture/arts/met-microsoft-mit-ai-open-access-hack/ https://www.microsoft.com/inculture/arts/met-microsoft-mit-a... There are other interesting applications e.g one that would presumably take an instagram feed and substitute images with "similar art" - https://www.microsoft.com/inculture/uploads/prod/2019/02/mylife-my-met.pdf https://www.microsoft.com/inculture/uploads/prod/2019/02/myl...
- rememberlenny 8y agoCode repo: https://github.com/Microsoft/GenStudio https://github.com/Microsoft/GenStudio
- mhamilton723 8y agoThanks for adding this!
- gambler 8y agoIt's good to see someone building something easily explorable. If people had more access to experiments like this, there would be far less BS hype around deep learning, since the public would get some intuitive feel for what those networks are doing, how they work and what are their limitations.
- gambler 8y agoBTW, this is a combination of several goblets and an armor: https://imgur.com/a/JPslbIs https://imgur.com/a/JPslbIs It find it hilarious how many people who ought to know better claim that GANs do not rely on memorization of source data. Obviously they do. At some point people even studied how connectivist memory could work (e.g. Hopfield networks, which I've just started looking into - fascinating stuff). Somehow that branch of research is never mentioned in any of the recent deep learning papers.
- mhamilton723 8y agoVery cool! You have left the realm of the MET Open Access data and ventured into the realm of the network that was pre-trained on imagenet!