DescendingNode's profile picture. space + robots + research

Kuldeep Barad

@DescendingNode

space + robots + research

Kuldeep Barad reposted

Yesterday we officially launched CAESAR: the Stanford Center for AEroSpace Autonomy Research: caesar.stanford.edu. Goal: solve the hard problems in spacecraft autonomy through the judicious incorporation of AI. More here: spacenews.com/stanford-cente… @StanfordEng @StanfordAILab


Kuldeep Barad reposted

As a researcher, I have two kinds of ideas: the bad ones, and those that other people had 5y ago.


Testing some of our 6-dof grasping models (ft. @billwurtz outro)


Kuldeep Barad reposted

... ... and here is how the Chandrayaan-3 Rover ramped down from the Lander to the Lunar surface.


Kuldeep Barad reposted

Large language models and web-scale data have some use in robotics as a user interface as nicely demonstrated here, but in my opinion they are not what we need to help with perception, object representation and precise planning which are the real current barriers in robotics.

Computers have long been great at complex tasks like analysing data, but not so great at simple tasks like recognizing & moving objects. With RT-2, we’re bridging that gap by helping robots interpret & interact with the world and be more useful to people. nytimes.com/2023/07/28/tec…



Super! For people in small and young ML teams, not having this acquired knowledge base is a huge barrier. Have been looking for this for a while.

Excited to announce our Deep Learning Tuning Playbook, a writeup of tips & tricks we employ when designing DL experiments. We use these techniques to deploy numerous large-scale model improvements and hope formalizing them helps the community do the same! github.com/google-researc…

zacharynado's tweet image. Excited to announce our Deep Learning Tuning Playbook, a writeup of tips & tricks we employ when designing DL experiments. We use these techniques to deploy numerous large-scale model improvements and hope formalizing them helps the community do the same! github.com/google-researc…


Its super weird to end the year with an incredible amount of learning and growth, but disproportionately high dissatisfaction at the same time. On the positive side, I have felt internal momentum building up to change this lately. So, the next year is full of optimism.


Kuldeep Barad reposted

Russel Kirsch was the inventor of the pixel. He passed away only a couple of years ago. The 172 x 172 image of his son made in 1957 is the 1st digital images ever made. It's a photo that changed the world. His legacy lives on in the billions of photos we take and share each day.

docmilanfar's tweet image. Russel Kirsch was the inventor of the pixel. He passed away only a couple of years ago. The 172 x 172 image of his son made in 1957 is the 1st digital images ever made. It's a photo that changed the world. His legacy lives on in the billions of photos we take and share each day.

Kuldeep Barad reposted

Blender visualization of a flexible camera model I made for rendering in #InstantNeRF so you can see what’s going on. Who says we have to use perspective cameras all the time?! I love NeRF so much. Cheap and easy photorealistic renders is a game changer for digital surrealism.


Kuldeep Barad reposted

Did you know @Meta has transitioned @PyTorch to the Linux Foundation? Read Mark Zuckerberg's post about the launch of the PyTorch Foundation here: hubs.la/Q01m2X8c0 #PyTorchFoundation #PyTorch #AI #OpenSource

linuxfoundation's tweet image. Did you know @Meta has transitioned @PyTorch to the Linux Foundation? 

Read Mark Zuckerberg's post about the launch of the PyTorch Foundation here: hubs.la/Q01m2X8c0
#PyTorchFoundation #PyTorch #AI #OpenSource

Getting out of a long slump with a Saturday ride in the gorgeous north of Luxembourg.

DescendingNode's tweet image. Getting out of a long slump with a Saturday ride in the gorgeous north of Luxembourg.
DescendingNode's tweet image. Getting out of a long slump with a Saturday ride in the gorgeous north of Luxembourg.
DescendingNode's tweet image. Getting out of a long slump with a Saturday ride in the gorgeous north of Luxembourg.
DescendingNode's tweet image. Getting out of a long slump with a Saturday ride in the gorgeous north of Luxembourg.

Kuldeep Barad reposted

"An Impartial Take to the CNN vs Transformer Robustness Contest" Are vision transformers really better than CNNs? This paper strongly suggests an answer, based on a robustness throwdown between {ViT, Swin} vs {BiT, ConvNeXt}. [1/10]

davisblalock's tweet image. "An Impartial Take to the CNN vs Transformer Robustness Contest"

Are vision transformers really better than CNNs? This paper strongly suggests an answer, based on a robustness throwdown between {ViT, Swin} vs {BiT, ConvNeXt}. [1/10]

Kuldeep Barad reposted

It takes Gandalf 77 years of "research" to verify that Bilbo's is the One Ring. And they tell me six months is a reasonable paper deadline, smh.


One of my biggest fears in life is to end up with such a narrow range of experiences that it makes one too opaque to absorb new perspectives, new knowledge, new stories and new people.


Kuldeep Barad reposted

Here is a closeup:


Kuldeep Barad reposted

I had to make a before and after to really appreciate how good the James Webb Telescope really is.


Kuldeep Barad reposted

BLOOM is here. The largest open-access multilingual language model ever. Read more about it or get it at bigscience.huggingface.co/blog/bloom hf.co/bigscience/blo…

BigscienceW's tweet image. BLOOM is here. The largest open-access multilingual language model ever. Read more about it or get it at
bigscience.huggingface.co/blog/bloom
hf.co/bigscience/blo…

Adding @weights_biases support hooks to the training stack has been so awesome and easy that I feel terrible for not doing it soon enough. What a slick platform !


Loading...

Something went wrong.


Something went wrong.