B3D AI Forum/Thread

I am surprised to see that there is no dedicated AI forum/thread on B3D. It's not really Graphics, Console, Mobile or PC yet seems to affect all of the above.

I noticed that, on the day they had the XBox Series X pre-orders, Microsoft also exclusively licensed OpenAI GPT-3 text generation.

https://www.theverge.com/2020/9/22/...-gpt-3-exclusive-license-ai-language-research

And since AI is now being used by Nvidia for DLSS and other interesting technologies, and Microsoft apparently using it in the Azure Cloud and what not, I had a look what Sony was up to and saw SonyAI Inc, founded April 1, 2020.

This being a technical forum, I wondered why no thread has been pinned for this extremely exciting technology. Maybe it's time?

https://azure.microsoft.com/en-us/
https://docs.microsoft.com/en-us/learn/azure/
https://www.nvidia.com/en-us/research/ai-playground/
https://www.nvidia.com/en-us/deep-learning-ai/
https://www.nvidia.com/en-us/industries/
https://www.ai.sony/
https://www.sony.net/SonyInfo/sony_ai/
https://neuralink.com/
https://machinelearning.apple.com/updates/introducing-aiml-residency-program
https://aws.amazon.com/ai/
https://blog.aboutamazon.com/amazon-ai
https://aws.amazon.com/machine-learning/what-is-ai/
https://ai.google/
https://experiments.withgoogle.com/ai
https://cloud.google.com/ai-platform/

Top Companies developing AI

https://www.datamation.com/artificial-intelligence/top-artificial-intelligence-companies.html
https://www.thomasnet.com/articles/top-suppliers/ai-software-companies/#register
 
It seems like there are enough people that dabble in it, I wouldn't mind having a sub forum for it if we want to dump all the AI related stuff into there.
 
Microsoft did some kind of exclusive licencing deal with openai about GPT-3. This could become super significant. GPT-3 is amazing at parsing text and understanding what it means. This can only get better and it might getter better very quickly

Unlike most AI systems which are designed for one use-case, OpenAI’s API today provides a general-purpose ‘text in, text out’ interface, allowing users to try it on virtually any English language task. GPT-3 is the most powerful model behind the API today, with 175 billion parameters,” OpenAI explains in a blog post about its partnership with Microsoft.

https://www.theverge.com/2020/9/22/...-gpt-3-exclusive-license-ai-language-research
 
Paging @Rys to weigh in on forum reorg for AI/ML focus.
 
some AI stuff i rememebr but got no links
* animating a photograph to follow a video (not just face but also full body animation)
* making songs from a few minutes of "source" song
* isolating various instruments and vocal from stereo audio
* generating new voice dialogues from a bunch of training voice (so you can fake dead artist for example)
 
Back
Top