"Neural codes of space" (https://t.co/dtWS8g35rq), very nice talk. If there is one thing that we know is that Kant was indeed right on many accounts for more than 240 years now.
I'm glad that we have great autonomous vehicles workshops in many conferences inviting folks from industry so that we provide the rare opportunity to share about everything they are not deploying today in their cars 😅
Giving a talk in the @Stanford SCIEN seminar this Wednesday (1/3) at 4:30pm:
https://t.co/EgDLBwv2YU
The topic is “normal coordinates”: a shape representation little-used outside of mathematics—but which turns out to have nice applications in geometry processing & learning.
One of the most amazing things I’ve ever seen: a standing ovation for the full Daraxonrasib results
I feel inspired and energised, to put it mildly — we have a targeted therapy for pancreatic cancer now, and nothing is undruggable anymore
Oxford, the longest running continuous weather station in UK history, with temperature observations stretching back to 1815, has preliminarily broken its maximum temperature record for May yesterday by OVER 3ºC with a temperature of 33.7ºC. Unprecedented in its 211-year history.
Introducing the newest Coral board, for efficient, on-device AI!
Check out the demos in the video:
- On-board speech translation
- Natural language controlling hardware
- Vision & sound generating music
The problem is that academia will spend the next year deciding who will be the first author of the paper and the model architecture, instead of just collecting the data.
It’s hard to see something that can have reasonable behaviors with so few parameters(<1B); and so much data scaling potential (1k hours of high quality mocap-retargeted data already gives a ton of capabilities). I really hope this time we can start with an open-source rat race towards humanoid foundation models, instead of letting companies develop everything internally. If history has told us anything, on early stages of a research field, open-source and joint forces are fundamental forces that makes it thrive.
In other words, “humanoid foundation models need foundational human efforts”. 3/3
I knew this was coming, the moment Gemini started to take a bit of the market share, there wouldn't be any other option other than curbing the compute to scale. This will backfire in their share for sure.
AI has now solved a major open problem -- one of the best known Erdos problems called the unit distance problem, one of Erdos's favourite questions and one that many mathematicians had tried.
https://t.co/SD1vVPkrHR
To MRL or not to MRL: Text Embeddings are Robust to Truncation Without Matryoshka Embeddings, Except In Heavy Truncation Scenarios
@sobamchan et al. show that random truncation of text embeddings is competitive with MRL.
📝 https://t.co/P8PfXxSSIs
👨🏽💻 https://t.co/l3w8oFHUlR