anyone know of any great comp neuro textbooks?

VQGAN + CLIP blue whales swimming through neon city prompts

VQGAN + CLIP "matte painting of Final Fantasy in the style of Akihiko Yoshida | trending on artstation" + 3D photo inpainting

Super excited to share something we have been working on for the last 1.5yrs. Check out our #RSS2021 paper on Rapid Motor Adaptation. RMA allows a legged robot trained *fully* in simulation to *adapt* online to diverse real-world terrains in real-time! ashish-kmr.github.io/rma-legged-rob…

I wonder if you could develop a cerebral organoid culture protocol + optical interface that did well on ImageNet. Surprisingly this doesn't seem to have been done yet?

from Tech: Demonstrating high fidelity mocap with CodeMiko

i want to get into ML research, what topic would you recommend? nothing. now it is not time to get into ML research. now its time to either observe what others are doing, or to build innovative applications using established techniques, or both.

Fine, I'll take the bait. This is actually the best time to get into ML research. Everything is so messy right now, and in that mess lies opportunities. There is so much to understand that if you have a passion for research, there's really no need to listen to much else.

So many exciting new frontiers in ML, it's hard to give a short list, particularly in new application areas (e.g. in the physical and biological sciences). But the Big Question is: "How could machines learn as efficiently as humans and animals?" This requires new paradigms.

he’s right from a talent allocation perspective from another perspective there’s quite a few problems researchers are not incentivized to explore

“Any idiot can build a bridge that stands, but it takes an engineer to build a bridge that barely stands.” - Unknown

This twirling work of science is special ✨ AFAIK, it's the first crystal structure of a functional #protein fully designed by #AI A milestone in our quest to use language models to generate proteins that are unseen in nature & can function well in the real-world. Read below👇

We are very impressed with how good #Alphafold is at predicting protein-protein interactions. A few tricks on the alignment, reranking models, and using the older model make a significant improvement. See our preprint at biorxiv.org/content/10.110…

Neural architecture search is computationally expensive, but this is prob one of those cases when it makes sense to use it, and first evolve a really efficient version of Transformer for next-step-prediction *before* training one of those 175Bn parameter large language models.

People who think evolution works through random mutations and selection need to explain how intelligent life appeared using nothing else. Clearly, any optimization process is more efficient if it uses some sort of gradient estimation.

Just published: Do large language models understand us? link.medium.com/0S1dajuU2lb It’s sometimes claimed that ML is "just stats" and AI can't "understand". I'm arguing that LLMs have a great deal to teach us about language, understanding, intelligence, sociality, even personhood.

Over the past year, life science is getting transformed by #AI: —Accurately predicting protein structure from amino acid sequence —Accurately predicting RNA structure —Step closer to predicting gene expression from DNA sequence (this week) nature.com/articles/s4158… Eeshit Dhaval Vaishnav Massachusetts Institute of Technology (MIT)

Deep learning works, symbolic models don't. It's that simple. If you want more symbolic models, then work hard and make them work. That's what NN people did, even when nobody believed in their research