Research Scientist at Google
Joined November 2014
- Tweets17
- Following30
- Followers347
- Likes58
Jonathan Tompson retweeted
Introducing 𝐀𝐋𝐎𝐇𝐀 𝐔𝐧𝐥𝐞𝐚𝐬𝐡𝐞𝐝 🌋 - Pushing the boundaries of dexterity with low-cost robots and AI. @GoogleDeepMind
Finally got to share some videos after a few months. Robots are fully autonomous filmed in one continuous shot. Enjoy!
Jonathan Tompson retweeted
For the past year we've been working on ALOHA Unleashed 🌋 @GoogleDeepmind - pushing the scale and dexterity of tasks on our ALOHA 2 fleet. Here is a thread with some of the coolest videos!
The first task is hanging a shirt on a hanger (autonomous 1x)
Jonathan Tompson retweeted
ALOHA 2 is here! 🤖
aloha-2.github.io/
The upgraded hardware has:
* More durability
* Greater fine-grained control
* Lower stress on operator
All off-the-shelf + open-source.
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!
Jonathan Tompson retweeted
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!
Jonathan Tompson retweeted
What happens when we train the largest vision-language model and add in robot experiences?
The result is PaLM-E 🌴🤖, a 562-billion parameter, general-purpose, embodied visual-language generalist - across robotics, vision, and language.
Website: palm-e.github.io
Jonathan Tompson retweeted
Digit Significance in Machine Learning by Vincent Vanhoucke in @TDataScience towardsdatascience.com/digit…
Jonathan Tompson retweeted
A summary of our recent work on the use of convex duality (specifically the use of Q-LP with convex regularization) in RL:
ai.googleblog.com/2020/07/du… via @googleai
Jonathan Tompson retweeted
This is an awesome explanation of our paper! Especially loved your description of the Temporal Self-similarity Matrix. You end up answering some questions that I was asked during the conference. Thanks for taking the time and effort @ykilcher
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
Jonathan Tompson retweeted
Welcome to Alphabet!
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.
Jonathan Tompson retweeted
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.
Jonathan Tompson retweeted
Check out RepNet, a single model that analyzes video to provide counting statistics and identify changes in patterns for a broad range of repeating processes—exercising, a bird flapping its wings, pendulums swinging & more. Read more and try it out at goo.gle/2CxtSkn
Jonathan Tompson retweeted
Introducing RepNet, a model that counts repetitions in videos of *any* action
w @yusufaytar, @JonathanTompson, @psermanet and Andrew Zisserman
Paper: openaccess.thecvf.com/conten…
Project: sites.google.com/repnet
Video: youtube.com/watch?v=bvqg-ngu…
#CVPR2020 #computervision #deeplearning
Jonathan Tompson retweeted
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.
Jonathan Tompson retweeted
@ikostrikov is presenting this at ICLR today! Join our zoom poster session to learn about ValueDICE -- a new & highly-performant off-policy imitation learning algo!
10am-12 & 1-3pm PST.
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.
Jonathan Tompson retweeted
How to scale-up multi-task learning?
Self-supervise plan representations from lots of cheap unlabeled play data (no RL was used).
learning-from-play.github.io
by @coreylynch, Mohi Khansari, @xiao_ted, @vikashplus, Jonathan Tompson, @svlevine and @psermanet
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/
Jonathan Tompson retweeted
Excited to share our work on self-supervised learning in videos. Our method, temporal cycle-consistency (TCC) learning, looks for similarities across videos to learn useful representations.#CVPR2019 #computervision
Video: youtube.com/watch?v=iWjjeMQm…
Webpage: sites.google.com/corp/view/t…