Took some time to update my website with jekyll these holidays, with dark mode and a functioning mobile view. Thanks Maruan Al-Shedivat for the really great template! Check it out here: akashsharma02.github.io

While I don't think fully volumetric geometry representations are the future of *rendering* (inefficiency), they have a beautiful property of well defined simplification and LODs. This great work davidlindell.com/publications/b… adds it naturally to coordinate networks / neural rendering.

It’s been 5 years since we launched @pytorch. It’s much bigger than we expected -- usage, contributors, funding. We’re blessed with success, but not perfect. A thread (mirrored at soumith.ch/posts/2022/01/… ) about some of the interesting decisions and pivots we’ve had to make 👇

Finally open-sourced my implementation of Code-NeRF. Code-NeRF is a latent variable auto-decoder that uses differentiable rendering to learn variations in shape and texture over a semantic category. Implemented with support for distributed training github.com/akashsharma02/…

iGone

Super excited to start my internship at Meta Reality labs today! I will be working with Tianwei Shen in Redmond on multi-modal fusion for localization and mapping!

This thread will be dedicated to the victims of the mass shooting in Uvalde, Texas. As a parent, my heart is broken seeing these faces and their ages. My condolences to their families during this difficult time 🙏🏻 Xavier Lopez, 10. He was a 4th grader at Robb Elementary.

Computer scientist Keenan Crane, PhD, is asked to explain fractals to 5 different people; a child, a teen, a college student, a grad student, and an expert.

Can we stop the “X is all you need” meme in ML. It literally says to stop thinking about everything else, and that’s just counterproductive.

The Feynman Razor Complexity and jargon are used to mask a lack of deep understanding. If you can’t explain it to a 5-year-old, you don’t really understand it. If someone uses a lot of complexity and jargon to explain something, they probably don’t understand it.

📜🚨📜🚨 NN loss landscapes are full of permutation symmetries, ie. swap any 2 units in a hidden layer. What does this mean for SGD? Is this practically useful? For the past 5 yrs these Qs have fascinated me. Today, I am ready to announce "Git Re-Basin"! arxiv.org/abs/2209.04836

So lucky to be working at FAIR right now! 💯💯💯

Our new work explores generating future observations (blue) given the past (gray), by leveraging large-scale pretraining of image diffusion models for video prediction, and conditioning on timestamps & invariant data modalities. w/ Deva Ramanan page: cs.cmu.edu/~tkhurana/dept…

To enable AI to physically interact with the world we are releasing research advances from our team AI at Meta & collaborators ai.meta.com/blog/fair-robo… More on 🎯Sparsh, ☝️Digit 360, 🦾Meta Digit Plexus, and 🤝partnerships w/ GelSight, Inc. & Wonik Robotics 🧵⬇ 1/

What a wonderful journey working on Sparsh. I am looking forward to seeing the community building and awesome research around tactile sensing that a shared backbone can unlock. We will be at #Corl2024 presenting our work. #Fulbright Fulbright Colombia

I'm in Munich @ #CoRL2024 and presenting 1. Sparsh - Nov 7 (morning session) sparsh-ssl.github.io 2. Digit 360 and Sparsh demo at the Meta booth (Nov 6 and 8: 3:00pm to 5:00pm) Looking forward to talk about state estimation and tactile perception for robotics!

For robot dexterity, a missing piece is general, robust perception. Our new Science Robotics work combines multimodal sensing with neural representations to perceive novel objects in-hand. 🎲 Featured on the cover of the November issue! #ScienceRoboticsResearch 🧵1/9

Excited to share that I'll be joining University of California at Irvine as a CS faculty in '25!🌟 Faculty apps: Krishna Murthy, Zhuang Liu & I share our tips: unnat.github.io/notes/Hidden_C… PhD apps: I'm looking for students in vision, robot learning, & AI4Science. Details👇