I'll be presenting my poster on learning from crowdsourced human feedback for robot learning #1404 at 5pm tomorrow (Thursday 14th) @NeurIPS Come say hi, and let's have some interesting conversations! #NeurIPS23

Robot learning in the real world can be expensive and unsafe in human-centric environments. Solution: Construct simulation on the fly and train in it! Excited to share RialTo, led by Marcel Torné on learning resilient policies via real-to-sim-to-real policy learning! A 🧵 (1/12)

Presenting a robotic system that works is messy environments and is robust to all kinds of variation — distractors, clutter, errors in execution. Check out more real-to-sim-to-real.github.io/RialTo/ and 👇

🥽 Want to use your new Apple Vision Pro to control your robot? Want to record how you navigate / manipulate the world to train a policy? I developed an app for VisionOS that can stream your head / wrist / finger movements over WiFi, which you can subscribe on any machines using

I really enjoyed reading Lucy Shi's latest work on incorporating language feedback in policy learning. Humans will be a key component for scaling up robot learning and this work shows a new and smart way for anyone to help their robots become more robust. 🤖🦾Awesome work!

Super excited to see more work on narrowing the sim-to-real gap by using real world data! Marius Memmel, Abhishek Gupta and co are using small amounts of real-world interactions to autonomously improve the simulator!🌎🤖

Check out Zoey Chen’s new work on automating real-to-sim transfer of articulated scenes! Super excited to see more people interested in the potential of real-to-sim-to-real transfer!🤖🌍

Today we’ll be presenting RialTo at #RSS2024 ! If you are around come to the presentation during the Manipulation session at 11am and let’s chat during the poster session later in the day! Looking forward to many exciting discussions!

A real-to-sim-to-real approach for more precise home robots. Full video: bit.ly/3WtzL8L

Thank you MIT CSAIL for covering RialTo!

An excellent piece by Stefan Milne covering some of our recent work pushing the paradigm of real-to-sim-to-real for scalable robot training, led by Zoey Chen Marcel Torné and many others across Allen School and MIT CSAIL! washington.edu/news/2024/08/0… Give it a read :)