Neural Networks
- Started
- Last post
- 46 Responses
- drgs2
Future of Photoshop
- uan1
Martin Ford's new book, 'Architects of Intelligence'
Interviews withs 23 of the most experienced AI and robotics researchers in the world, asking them about the current state of AI, how AI can be applied to solve useful problems, and what that means for the future of robotics and computing.
- uan1
Music Transformer: Generating Music with Long-Term Structure
- drgs1
Text to image!
Try with "apple" to see if it works. Responds to very few words, it seems
http://t2i.cvalenzuelab.com/
- uan1
In new computer experiments, artificial-intelligence algorithms can tell the future of chaotic systems. They hacked the 'butterfly effect'
- drgs1
- drgs1
- drgs1
- uan1
- drgs3
- drgs3
The strongest chess engine so far was Stockfish (ELO 3400, highest human rating = ELO 2850 or so), which basically calculates all possible moves at a certain depth.
Deep Mind's Alpha Zero, which learned chess in 4 hours by playing with itself, beats Stockfish with crazy moves and sacrifices that no one understands.
This is a final goodbye to humans in chess
- uan0
DeepSpeech
Mozilla’s Open Source Speech Recognition Model and Voice Datasethttps://github.com/mozilla/DeepS…
Project DeepSpeech is an open source Speech-To-Text engine. It uses a model trained by machine learning techniques, based on Baidu's Deep Speech research paper. Project DeepSpeech uses Google's TensorFlow project to make the implementation easier.
- drgs2
"Coditany of Timeness” is a convincing lo-fi black metal album, complete with atmospheric interludes, tremolo guitar, frantic blast beats and screeching vocals.
It was generated by two musical technologists using a deep learning software that ingests a musical album, processes it, and spits out an imitation of its style.Goosebumps...
https://dadabots.bandcamp.com/al…
https://theoutline.com/post/2556…
- i_monk1
I've been wondering whether it'd be possible to take a collection of recorded speech – say David Attenborough's narration of dozens of documentaries – and apply a neural network to it to synthesize completely new speech. I've seen attempts to produce music this way, working from samples of classical pieces, but nothing worthwhile emerges.
- already doing it https://arstechnica.…scarabin_net
- scary cooli_monk
- sofas-1
Building neural networks for fun/work is acceptable (many times encouraged) because we can hardly empathize with them and because many times we make humans for the same reasons, so it seems normal.
I'm not totally against it, but this practice isn't really criticised.
- They aren't people.monospaced
- Just fancy calculators.monospaced
- true (even though we're talking *extremely* fancy calculators). Still, we're heading towards beings having a consciousness on par if not higher than humanssofas
- That's debatable, and a well documented concept.monospaced
- yup.
also sorry for this coming out quite provocative.sofas - No apologies necessary.monospaced
- uan0
- drgs1
- drgs1