Go to bottom

A.I. related stuff

category: code [glöplog]

"Training Humans, conceived by Kate Crawford, AI researcher and professor, and Trevor Paglen, artist and researcher, is the first major photography exhibition devoted to training images: the collections of photos used by scientists to train artificial intelligence (AI) systems in how to “see” and categorize the world."
added on the 2019-09-18 20:34:35 by wertstahl wertstahl
Neural Composer generates video game music using neural networks. (Can save as .mid !)

https://www.youtube.com/watch?v=UWxfnNXlVy8 (Demo)
added on the 2019-09-24 23:21:55 by wertstahl wertstahl
This has surprisingly little attention, but it is a huge warning shot towards Hollywood:

https://www.youtube.com/channel/UCKpH0CKltc73e4wh0_pgL3g Ctrl Shift Face
Contains very intense "Deepfaking" and also Voiceswap demos.
added on the 2019-09-24 23:25:14 by wertstahl wertstahl
See how a modern neural network completes your text. Type a custom snippet or try one of the examples.
added on the 2019-09-25 01:44:31 by wertstahl wertstahl
Self driving car simulator that uses the Unreal Engine.
added on the 2019-09-25 22:56:48 by neoneye neoneye
talktotransformer, lol. impressive, but you can confuse it with highly specific and/or weird stuff.

stopped messing with it at "The beauty of Eulers identity is more erotic "
added on the 2019-09-26 22:58:35 by jco jco
"But what if slugs could talk?" worked well though.
added on the 2019-09-26 23:01:28 by jco jco
singularity is still far away it seems...

BB Image
The question is always: beyond the demo, what are possible usecases and what will be the impact on our reality. (Video semi related) https://www.youtube.com/watch?v=SjsTzWXPBLM
But lets keep dumping related stuffs in here, i am sure, one day, Narrow Artificial Intelligence (NAI) will even find it's way into the first demoscene product.
added on the 2019-09-27 04:37:16 by wertstahl wertstahl
@wertstahl, i could link you at least one real usecase for environmental monitoring (here) where i am involved on the scientific side. To put it short, reality is brutal. Images are of bad quality, noisy, most of the times not annotated properly, and the real work load has mostly nothing to with AI. It is general programming, datamanagement, data provenance, work flows, organisation, system installation and monitoring. AI frameworks, net structures, drivers and accelerators come basically for free. So if AI is part of a solution than mostly it's exactly that, a part of it.

I thought about putting some kind of autoencoder into a tiny intro (the decoding part) with some clever system of shared and reused filters, but so far that's just an idea ;)

Until the AI overlords arise and bring doom to as all i will happily enjoy this dump. The Better Call Trump thingy was neat =)
added on the 2019-09-27 09:57:28 by HellMood HellMood
Interesting linked content, thanks.

A "more accessible" entertaining example for people on the depth/complexity of planning and social engineering possible by A.I.:
added on the 2019-09-28 22:12:45 by Hoild Hoild
Not sure if AI is involved, but here is a tool that automatically generates mixes from uploaded parts, parts can be organized in packs and assigned to channels and the main genre channel then fuses everything together into one endless stream. No more artists required.

https://mubert.com/ (free tryout for producers in the producers tab or beta.mubert.com )

Example Mix Channel
added on the 2019-10-01 22:18:40 by wertstahl wertstahl
Some press, covering Mubert. https://www.musicbusinessworldwide.com/ai-in-the-music-industry-is-nothing-to-be -scared-of-its-a-step-towards-an-exciting-new-future/

All in all, AI is a game changer due to one very important thing: from now on, music can adapt perfectly to a listener; listeners no longer have to adapt to music.
added on the 2019-10-01 23:13:03 by wertstahl wertstahl

Oh man. Sick minds are sick.


Capable of analysis on target personnel's sex (male, female), ethnicity (such as Uyghurs, Han) and color of skin (such as white, yellow, or black), whether the target person wears glasses, masks, caps, or whether he has beard, with an accuracy rate of no less than 90%.
added on the 2019-11-12 03:10:26 by wertstahl wertstahl
Yeah, that's why the Chinese are investing so much in AI. They want total surveillance of ethnic minorities.
added on the 2019-11-12 08:24:05 by Adok Adok
Related: Usecase


Spleeter is the Deezer source separation library with pretrained models written in Python and uses Tensorflow. It makes it easy to train source separation model (assuming you have a dataset of isolated sources), and provides already trained state of the art model for performing various flavour of separation :

Vocals (singing voice) / accompaniment separation (2 stems)
Vocals / drums / bass / other separation (4 stems)
Vocals / drums / bass / piano / other separation (5 stems)
added on the 2019-11-17 04:11:54 by wertstahl wertstahl
Semi-Related and actually very ahead of it's time


Neural Network on a Commodore 64

-- in BASIC
added on the 2019-11-22 21:58:05 by wertstahl wertstahl
The obsolescence of human Musicians advances one step at a time. All the time.



links provided by dadabots.
added on the 2020-05-06 00:18:53 by wertstahl wertstahl

The moment has arrived that you better only talk IRL face to face, or it might be ... "questionable" who and what about you are talking to, in terms of who has control over the outcome of the conversation.
added on the 2020-05-22 07:20:15 by wertstahl wertstahl
Proprietary commercial Software for upscaling Videos. Just for usecase scenarios. I am not affiliated with Topaz Labs, i don't recommend this software. I think it is just good to know about it.
added on the 2020-07-10 19:34:33 by wertstahl wertstahl
BB Image

GPT3 generated text for Kayne West 2020 presidential campaign.

See Arram Sabeti GPT3 article for more hillarious generated texts.
added on the 2020-07-11 09:45:08 by neoneye neoneye


Go to top