Imagine being an international student, you finished all your courses & you're getting close to graduation: 1. Your OPT (gives you a temp. work permit) will likely be cancelled. 2. H1B can be cancelled anytime too. 3. You can't really go home because of the pandemic. And now:

Thrilled to have graduated my first PhD student today! Daniel did fantastic work at UT, and I can't wait to see what he does during his postdoc at Berkeley.

Announcing the robot learning seminar series at milamontreal and REAL - Robotics and Embodied AI Lab. Great speaker lineup! See montrealrobotics.ca/robotlearnings… for schedule. First talk this Friday (i.e., tomorrow/today) by Florian Shkurti!

Here's a talk by Scott Niekum at CMU Robotics Institute yesterday on how to make inverse reinforcement learning safe and efficient - very exciting works and lots of super cool insights! youtube.com/watch?v=sNByxR…

The NumPy paper is out! nature.com/articles/s4158…

Hey, I got tenure.

I successfully defended my PhD! 👨‍🎓😃 Thanks a lot to my advisors Scott Niekum and Ashish, and my committee Farshid Alambeigi, PhD, Raul Longoria and Dongmei for their tremendous support and help!

Excited to share our new work on learning distributions over object articulation models from depth images: openreview.net/forum?id=H1-uw… just got accepted at #CoRL2021. Joint work with Stephen Giguere, Rudolf Lioutikov, and Scott Niekum . #CoRL #robotics #robots #learning #perception

Booked a flight in April, San Francisco to Bhopal and cancelled. @airindiain hasn't refunded the money yet. Have filled 3 feedback forms, and multiple emails. No response yet. Jyotiraditya M. Scindia PMO India

This week at CoRL: Inaccurate point estimates of articulated object parameters can lead to failed manipulation. We introduce DUST-net, which can infer distributions over articulation model parameters from raw depth images using a representation based on screw theory. 1/

See our poster on Thurs (session 8) and full paper for more details: arxiv.org/abs/2108.05875 w/ Ajinkya Jain 2/n

Our systems streamline complex kitting programs and introduce greater flexibility while packaging items into virtually any container. Is your organization ready to level up their kitting operation? #robotics #automation

RT-X: generalist AI models lead to 50% improvement over RT-1 and 3x improvement over RT-2, our previous best models. 🔥🥳🧵 Project website: robotics-transformer-x.github.io

We can tell our robots what we want them to do, but language can be underspecified. Goal images are worth 1,000 words, but can be overspecified. Hand-drawn sketches are a happy medium for communicating goals to robots! 🤖✏️Introducing RT-Sketch: rt-sketch.github.io 🧵1/11

“Place the coke can on the left” can mean the upper left corner, the bottom left, etc. A goal image can address this, but is too environment-specific to apply to new tablecloths or kitchens, for instance. What lets us convey goals minimally yet expressively? Sketches! 🧵2/11