Jonathan Tompson

Led by @GoogleDeepMind, we present ALOHA 2 🤙: An Enhanced Low-Cost Hardware for Bimanual Teleoperation. ALOHA 2 🤙 significantly improves the durability of the original ALOHA 🏖️, enabling fleet-scale data collection on more complex tasks. As usual, everything is open-sourced!

New Paper Explained Video 🥳 This network learns end-to-end how to count repeating actions in video sequences, even when they speed up, slow down or only happen in a part of the video. youtu.be/qSArFEIoSbo @GoogleAI @debidatta @yusufaytar @JonathanTompson @psermanet

Welcome @Rob_Fergus to @DeepMind. He's a computer vision and deep learning pioneer & a leading scientist in AI. I’m delighted he’s joined @DeepMind (remaining Prof @ NYU), looking forward to working with him again many years after we were at NYU.

arxiv.org/abs/1912.05032 <-- my most recent paper with @ikostrikov & J. Tompson introducing *ValueDICE* - an off-policy imitation learning alg. We set new SOTA for online imitation learning, and for the 1st time (afaik) beat behavior cloning in the totally offline regime.

Excited to share our new work on learning from play! We show a single agent, after self-supervising on 3 hours of play data, can generalize to 18 zero-shot manipulation tasks with 85% success. interactive paper: learning-from-play.github.io 1/