Life update: I’m fortunate to start as Applied Scientist at The Boston Dynamics AI Institute this week! I'm excited to work on some great problems with some great people.
https://t.co/EvFI9WbdZT
Please stop making new words for simple things to make it sound fancy!!!
Neural Physics == Model based method that have been used for ages with all kinds of optimizers RL IL trajOpt Evolutionary methods, sampling based search, etc.
Excited to share recent work with @kaihuac5 and @RamananDeva where we learn to do novel view synthesis for dynamic scenes in a self-supervised manner, only from 2D videos!
webpage: https://t.co/e5duAXm1lm
arxiv: https://t.co/Sh98RtFaJA
code (soon): https://t.co/V4JwF4v7po
https://t.co/NLv72CGl6W Many robotics researchers today hold a pessimistic view and think manipulation is solved and the only thing left to do is to scale up. This is why I really like this article, which shows a few potentially very large gap that most people have ignored.
Excited to share that our paper, "Multi-Robot Motion Planning with Diffusion Models," has been selected for a Spotlight presentation at #ICLR2025!🔦🤖
We scale diffusion planning to dozens of robots *without* multi-robot data by using search.
Project: https://t.co/NZk8rEVW2t 🧵
🎇 I’m on the academic job market! I’m a PhD candidate at @mldcmu.
My research tackles challenges that arise from the sequential nature of human-AI interaction. Toward this goal, my work involves:
🤖 reinforcement learning,
🧠 foundation models, and
👩💻 human-centered AI.
Shares appreciated!
Jack Campanella and Jillian Mallon made everything that went into hosting me look effortless, and @roboticseabass and Drew Hoelscher are awesome coworkers who helped me with a few ideas for my talk and abstract last minute. [5/n; n=5]
Last week, I had a great time visiting and speaking for the @GRASPlab's GRASP SFI Seminar at @Penn!
My talk included a mix of technical and non-technical topics. [1/n]
I also had insightful one-on-one conversations with graduate students, postdocs, and professors—the sense of collaboration and enthusiasm at @GRASPlab I'd heard so much about was evident right away. [4/n]
We are lucky for all the free, public, accessible, and open-source software/tutorials/presentations @roboticseabass consistently produces. I don’t know how he does it. And I’m extra lucky to also be his coworker 😛
Do check out his work if you haven’t already!
As an academic, I’m clearly biased. Yet, seeing the Open AI drama unfold, I can’t help but think that, for all their flaws, universities are still the best way to keep knowledge creation open.
“The thing I’m most proud of in my work are the people.” “At the end of the day, we’re in a people business and seeeing them succeed is really the biggest joy that I get.”