The model achieved an ImageNet 64x64 FID of 6.74.
🧠 Today we introduce Un-0 from @unconvAI : the first large-scale generative model build on physics as a compute primitive. This represents a “hello world” moment for physics-based models. We use the inherent time-varying behavior of physical systems to do compute for us. The result is a new way to build a computer that can be VASTLY more power efficient. 🧵 https://unconv.ai/blog/introducing-un-0-generating-images-with-coupled-oscillators/
Many users praised UnconvAI's Un-0 as a brilliant physics-based image generator using coupled oscillators because it offers a novel, brain-like compute paradigm they see as the future.
No Digg Deeper questions have been answered for this story yet.
#1267The official thread from @unconvai !
Introducing our first model, Un-0!
We trained an image generator powered by a backbone of coupled oscillators in place of a more traditional conventional neural network.
Did you know this wiggly thing can be a computer?
We think of computation as symbolic: add, subtract, multiply....
…but…you can also compute directly with voltage, current, charge, waves, materials (octopus arms…look it up)...or wiggles.
Today, we’re (@unconvAI ) sharing Un-0: our first step toward scaling this idea to modern AI workloads.
🧠 Today we introduce Un-0 from @unconvAI : the first large-scale generative model build on physics as a compute primitive. This represents a “hello world” moment for physics-based models. We use the inherent time-varying behavior of physical systems to do compute for us. The result is a new way to build a computer that can be VASTLY more power efficient. 🧵 https://unconv.ai/blog/introducing-un-0-generating-images-with-coupled-oscillators/
We’re at the beginning of the unconventional AI journey, but we are releasing artifacts to the community to enable us all to build together. We’ve open-sourced the Un-0 weights, training scripts, and ablation code.
If you build physics-based models, plug them into our scaffold and share your success!
Read the blog: https://unconv.ai/blog/introducing-un-0-generating-images-with-coupled-oscillators
Visit our GitHub: https://github.com/unconv-ai/Un-0
Introducing our first model, Un-0!
We trained an image generator powered by a backbone of coupled oscillators in place of a more traditional conventional neural network.
How does it work? Un-0 uses a system of coupled Kuramoto oscillators.
Initial random phases evolve under the pull of other oscillators through learned coupling strengths.
Is the physics actually doing the work? Yes.
Our ablations show that the trained dynamics provide value over a decoder only baseline and a random Kuramoto feature reservoir, and that increasing the number of integration steps increases model quality.
blog: https://unconv.ai/blog/introducing-un-0-generating-images-with-coupled-oscillators/
repo: train or extend it yourself. https://github.com/unconv-ai/Un-0
Did you know this wiggly thing can be a computer?
We think of computation as symbolic: add, subtract, multiply....
…but…you can also compute directly with voltage, current, charge, waves, materials (octopus arms…look it up)...or wiggles.
Today, we’re (@unconvAI ) sharing Un-0: our first step toward scaling this idea to modern AI workloads.
The results? We scale Un-0 up to 16k oscillators and 322M parameters and achieve an FID of ~6.74 on ImageNet 64x64.
To our knowledge, this is the most capable model based on a simulation of a physical system to date, expanding the Pareto frontier for small generative models. Though there is still remaining work to scale model quality as a function of model size.
At Unconventional AI, we're building a new kind of computer that runs AI on the dynamics of a physical system, at a fraction of the energy today's machines need.
We’re leveraging the dynamics of a physical system, such as the noisy, time-varying behavior of analog circuits that compute with analog voltage and current instead of conventional digitized numbers.
Un-0 illuminates a path towards running modern AI workloads on physical substrates that are more efficient than today's hardware.
with only 64 spins, we can't run ImageNet, but we can generate some very nice images of squares!
https://github.com/zachbe/digial-ising
Introducing our first model, Un-0!
We trained an image generator powered by a backbone of coupled oscillators in place of a more traditional conventional neural network.
@NaveenGRao @unconvAI game on brother.
🧠 Today we introduce Un-0 from @unconvAI : the first large-scale generative model build on physics as a compute primitive. This represents a “hello world” moment for physics-based models. We use the inherent time-varying behavior of physical systems to do compute for us. The result is a new way to build a computer that can be VASTLY more power efficient. 🧵 https://unconv.ai/blog/introducing-un-0-generating-images-with-coupled-oscillators/
How does this relate to energy efficiency? Most energy in existing von Neumann machines goes into moving information between memory and compute elements. Dynamical systems combine compute and memory into a single entity. What’s more, dynamical systems can tolerate noise. This opens up new opportunities to save energy in communication even further.
Why is this significant? It shows that computing is not some unique invention from humans; it’s present all through nature and physics. All physics of all physical entities has time; however, current computing systems don’t. We’re exploiting that time dimension.
Why is this significant? It shows that computing is not some unique invention from humans; it’s present all through nature and physics. All physics of all physical entities has time; however, current computing systems don’t. We’re exploiting that time dimension.
And the official @unconvAI post!
Introducing our first model, Un-0!
We trained an image generator powered by a backbone of coupled oscillators in place of a more traditional conventional neural network.
Churchill quote that works pretty well here (h/t CFO Ali) ... "Now this is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning." We are embarking on a fantastic journey 🚀
Un-0 represents a big first step in changing the paradigm of compute to dynamical systems. We are connecting intelligence to dynamics with this model release. Dynamics are a natural framing for compute for AI; neural networks themselves are really dynamical systems so the mapping becomes more straightforward. The brain does not have an abstraction of linear algebra; so in effect, we’re cutting out the middleman.
How does this relate to energy efficiency? Most energy in existing von Neumann machines goes into moving information between memory and compute elements. Dynamical systems combine compute and memory into a single entity. What’s more, dynamical systems can tolerate noise. This opens up new opportunities to save energy in communication even further.
@NaveenGRao @unconvAI How do you do credit assignment during training?
@davidcox i actually made a proof of concept for hopfield networks :)
https://www.zach.be/p/implementing-hopfield-networks-on
