🔥Check out our latest #NeurIPS2025 paper on end-to-end learning *Parameterized Skills* from demonstrations! We show that learning discrete skills with continuous parameters to refine their behavior can lead to better generalization on downstream tasks especially OOD tasks.

Wow, demos and the amount of data are seriously impressive🤩! I was just wondering, is it a good idea to use the MSE validation loss to prove the “scaling law”?

Stop Animal Abuse!!

🧵 Evaluating robot policies in the real world is slow, expensive, and hard to scale. During my internship at SceniX AI this summer, we had many discussions around the two key questions: how accurate must a simulator be for evaluation to be meaningful, and how do we get there?

I will join Northwestern University Computer Science as an Assistant Professor in Fall 2026! I am actively recruiting PhD students and seeking collaborations in robotics, human-robot interaction, brain-computer interfaces, cognitive science, societal impact of AI & automation,

This is how we use human videos for #RobotManipulation Present HRT1: One-Shot Human-to-Robot Trajectory Transfer irvlutd.github.io/HRT1/ 1️⃣ Collect demos 2️⃣ Extract 3D hand motion 3️⃣ Transfer to robot 4️⃣ Optimize base & trajectory ✅ One-shot imitation for mobile manipulation👇

Contact reasoning is all we need for manipulation, locomotion... Excited to see this work, which uses a learned ego-vision world model to reason around "good" contacts and "bad" contacts.

.Yunzhu Li is looking for PhD students interested in robot learning to work in his lab. To find out more about him - yunzhuli.github.io. For info on our #computerscience PhD programs bit.ly/CSPhDprogram. The deadline to apply is December 15.

How do we fuse VLM physical intuition with runtime adaptation in robot manipulation? It requires careful uncertainty quantification to properly weight the two. This is the key behind Maggie Wang 's Phys2Real approach. Congrats Maggie and team on this insightful work!

From a positive side, these papers are always looking for more challenging tasks. 🤣

Excited to see the GigaHands dataset (Rao Fu) collected using our BRICS capture system (Srinath Sridhar) being used for dexterous manipulation in the new SPIDER paper! It's a promising physics-based retargeting framework that uses large-scale sampling with virtual

Congrats to the team for their hard-core robotics work from RAI Institute!

Bro...

We are delighted to host Yunzhu Li from Columbia University this Friday, presenting "Foundation Models for Robotic Manipulation: Opportunities and Challenges". More details: events.brown.edu/computer-scien…

Robotics version: I have met many students who claim to be working on robotics but do not even know the basics of kinematics or motion planning.

I will join UChicago CS UChicagoCS as an Assistant Professor in late 2026, and I’m recruiting PhD students in this cycle (2025 - 2026). My research focuses on AI & Robotics - including dexterous manipulation, humanoids, tactile sensing, learning from human videos, robot

Huge congrats to Yiqing Liang. Witnessing your journey (from the first three papers that kept getting rejected, to that fourth one winning a Best Paper Nomination at CVPR) has been truly inspiring. All the best at Luma AI! I'll never forget the uncountable hotpot (our first

It’s disappointing to see the largest consumer product company, Apple, making such a preventable mistake. This lack of rigor may have already spread throughout the company, as their products now have much more bugs than before.