Now that #NeuralNetworks have had repeated big successes over the last 15 years, we are starting to look for better ways to implement them. Some new ones for me:

#Groq notes that NNs are bandwidth-bound from memory to GPU. They built a LPU specifically designed for #LLMs
https://groq.com/

A wild one — exchange the silicon for moving parts, good old Newtonian physics. Dramatic drop in power utilization and maps to most NN architectures (h/t @FMarquardtGroup)

https://idw-online.de/de/news820323

The news release on our recent publication about physical self-learning machines with a new all-physical training procedure, explained in layperson's terms:

Efficient training for artificial intelligence https://idw-online.de/de/news820323

#neuromorphic #NeuralNetworks
@MPI_ScienceOfLight

Obsidian, Zettelkasten and productivity. I've been looking for the perfect note-taking app for a very long time. As a result, my choice is a markdown, connections between notes and complete freedom of writing.

I fell in love with Obsidian at first sight and now I have my own knowledge base, a second brain.

https://obsidian.md/

#obsidian #zettelkasten #Productivity #neuralnetworks

Hi #EconTwitter!

On Friday, Sep 8 at 11AM EDT/4PM GMT/5PM CEST, we will resume the summer break by hosting another great #AMLEDS #Webinar with Lasse Heje Pedersen (Copenhagen Business School) talking about "#Machinelearning and the Implementable Efficient Frontier".

Andrew Chen (Federal Reserve Board) will moderate.

Link to paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4187217

If you are interested, please join us and register at https://lnkd.in/egTakd2

#finance #portfolio #investment #NeuralNetworks #BigData

Rethinking open source for AI

Check it out! 👇
https://www.infoworld.com/article/3706091/rethinking-open-source-for-ai.html#tk.rss_all

#SoftwareLicensing #NeuralNetworks #ArtificialIntelligence #OpenSource

Rethinking open source for AI - We keep using the term “open source” in the context of large language models (LLMs) li... - https://www.infoworld.com/article/3706091/rethinking-open-source-for-ai.html#tk.rss_all #artificialintelligence #softwarelicensing #neuralnetworks #opensource

Looking forward to our workshop 'Frontiers of Neuromorphic Computing' that will start tomorrow morning at the @MPI_ScienceOfLight Max Planck Institute for the Science of Light.

We are trying to record at least most of the talks and release them later (with some delay) on video.

https://indico.mpl.mpg.de/event/15/

#neuromorphic #NeuralNetworks #machinelearning

Watching Neural Networks Learn:

https://www.youtube.com/watch?v=TkwXa7Cvfr8&t=72

#NeuralNetworks #Learning #Functions

Amazing video!

Make Use Of: 3 Ways NVIDIA's Neuralangelo Can Benefit Creators https://www.makeuseof.com/nvidia-neuralangelo-benefits-for-creators/ #Tech #MakeUseOf #TechNews #IT via @morganeogerbc #ArtificialIntelligence #NeuralNetworks #3DModeling #Creative #Nvidia

Is there some thePyrateBay for #NeuralNetworks? Want to test #CodeLlama but Meta require some jumps over hoop to download it

#LLM

Neural quantum states utilize the impressive power of artificial #neuralnetworks to encode the fundamental object in #quantummechanics - the quantum many-body wave function.

What has remained a key open question is to understand for which #quantum states this encoding is efficient and therefore have low complexity. In a recent work we show that mean-field theories are simple in this context:

https://arxiv.org/abs/2308.10934

Wired: AI Can’t Read Books. It’s Reviewing Them Anyway https://www.wired.com/story/plaintext-ai-chatgpt-book-reviews/ #Tech #wired #TechNews #IT #Technology via @morganeogerbc #artificialintelligence #Business/TechCulture #neuralnetworks #deeplearning #Bookreviews #Plaintext #Business #ChatGPT #Books

Wired: AI Can’t Read Books. It’s Reviewing Them Anyway https://www.wired.com/story/plaintext-ai-chatgpt-book-reviews/ #artificialintelligence #Business/TechCulture #neuralnetworks #deeplearning #Bookreviews #Plaintext #Business #ChatGPT #Books

@m @rotopenguin @davidgerard I've always found it fascinating / sad that #neuralnetworks are worst at what computers were best at (counting, retaining and recalling minute information, easy lookups). Would be nice for #AI to find a way to merge the two capabilities

Loved this post, via @emilymbender, about "AI", #language, #ableism and more - flagging for @lingthusiasm as I think you'll enjoy it too. Big recommend to everyone though.

https://karawynn.substack.com/p/language-is-a-poor-heuristic-for

#autism #deaf #AI #LLM #NeuralNetworks #communication

A new type of #neuralnetworks and #AI 3/3

...machine is used to make a better one.

I've been thinking of a suitable set of tasks for these systems to optimize on and developed one earlier in #C. Recently I've come across the #micromouse competitions, and a #maze that a system needs to navigate and learn to go through it as fast as possible may be a good choice.

If anyone is interested in #collaboration on any of these aspects, do let me know!

A new type of #neuralnetworks and #AI 2/3

... Recent successes at one shot tasks where states changes become learning also point in this direction.

Free neural networks from being #feedforward. Genetic algorithms come to mind for training, but we'd need a suitable base structure that supports gene exchanges & some kind of recursion for repeated structures.

Another idea is to use backpropagation to develop a new learning method, just like when a #programming language or manufacturing...

A new type of #neuralnetworks and #AI 1/3

I've been thinking that #backpropagation based neural networks will reach their peak (if they haven't already), and it may be interesting to search for a new learning method. Some observations and ideas:

The two main modes of #neuralnetworks - training when weights are adjusted, and prediction when states change should be merged. After all real-life brains do prediction and learning at the same time, and they are not restarted for every task. ...

A new type of #neuralnetworks and #AI

I've been thinking that #backpropagation based neural networks will reach their peak (if they haven't already), and it may be interesting to search for an alternative #machinelearning method. I've had some observations and ideas:

The two main modes of #neuralnetworks - training when weights are adjusted, and prediction when states are adjusted should be merged. After all real-life brains do prediction and learning at the same time, and they survive for long-term; they are not restarted for every task. Recent successes at one shot tasks where states changes effectively become learning also point in this direction.

It'd also be nice to free neural networks from being #feedforward, but we don't have any training methods for freer networks. Genetic algorithms come to mind, but a suitable base structure has to be found that supports gene exchanges to some kind of recursion for repeated structures. Even with these it's unclear if we'd get any results in a reasonable amount of time.

Another idea is to use the current, imperfect learning method (backpropagation) to develop a new one, just like when #programming language or manufacturing machine can be used to make a new, better one. Here #backprop would be used to learn a new learning method.

I've been thinking if suitable playgrounds (set of tasks) for these systems to operate in and developed one earlier in #C. Recently I've come across the #micromice competitions, and a #maze that a system needs to navigate and learn to go through them as fast as possible may also be an interesting choice.

If anyone is interested in #collab #collaboration on any of these aspects, even just exchanging thoughts, let me know!

Wired: Geoffrey Hinton, Godfather of AI, Has a Hopeful Plan for Keeping Future AI Friendly https://www.wired.com/story/plaintext-geoffrey-hinton-godfather-of-ai-future-ai/ #Tech #wired #TechNews #IT #Technology via @morganeogerbc #artificialintelligence #Business/TechCulture #machinelearning #neuralnetworks #deeplearning #Neuroscience #Plaintext #Business