Toru (@ToruO

Toru

144 posts

Toru

@ToruO_O

mts @Amazon Frontier AI & Robotics prev: ai phd @UCBerkeley & researcher @meta @NVIDIA, cs undergrad @MIT, utokyo dropout 🌈 she/her/hers I like capybaras :D

Joined October 2020

Pinned
Toru
@ToruO_O
Mar 4
Peeling a potato is trivial -- until you try to make a robot do it with a knife. This is actually one of the hardest problems in manipulation: contact-rich, force-sensitive, and success is subjective. We taught a robot arm to peel with >90% success :D toruowo.github.io/peel
00:00
32K
Toru
@ToruO_O
Feb 28, 2025
Sim2Real RL for Vision-Based Dexterous Manipulation on Humanoids toruowo.github.io/recipe/ TLDR - we train a humanoid robot with two multifingered hands to perform a range of dexterous manipulation tasks robust generalization and high performance without human demonstration :D
00:00
50K
Toru
@ToruO_O
Apr 26, 2024
Imitation learning works™ – but you need good data 🥹 How to get high-quality visuotactile demos from a bimanual robot with multifingered hands, and learn smooth policies? Check our new work “Learning Visuotactile Skills with Two Multifingered Hands”! 🙌 toruowo.github.io/hato/
00:00
61K
Toru
@ToruO_O
Mar 5, 2024
Achieving bimanual dexterity with RL + Sim2Real! toruowo.github.io/bimanual-twist/ TLDR - We train two robot hands to twist bottle lids using deep RL followed by sim-to-real. A single policy trained with simple simulated bottles can generalize to drastically different real-world objects.
00:00
62K
Toru
@ToruO_O
Apr 5, 2023
It’s year 2023 and PPO is still going strong 🙂 open-sourcing a simple & performant PPO that trains Cartpole in <1m (with IsaacGym) github.com/ToruOwO/minima… same power as IsaacGymEnvs' default but everything in 4 files / under 1k lines of code
GitHub - ToruOwO/minimal-stable-PPO: A minimal and stable PPO.
From github.com
15K
Toru
@ToruO_O
May 24, 2024
A common question we get for HATO (toruowo.github.io/hato) is: can it be more dexterous? Yes! The first iteration of our system actually achieves this -- by capturing finger poses with mocap gloves and remapping them to robot hands. [video taken in late 2023 (with @YuZhang)]
00:00
00:00
64K
Toru
@ToruO_O
Oct 31, 2023
A fun side project from a while back : D Idea: teleop a dexterous hand with $5 flex sensors could be improved via better sensors, exoskeleton structures, etc... not following up but if anyone's interested, happy to share details & help! (with @brenthyi & Ting-Hao Wang)
00:00
8.7K
Toru
@ToruO_O
Jan 6, 2025
A great pleasure working with @JitendraMalikCV on the course! I'm always inspired by Jitendra's genuine passion for staying at the cutting edge of developments, and the abilities to new learn things so fast!! Highly recommend this unique and fun robotics course to everyone!!!
Jitendra MALIK
@JitendraMalikCV
Jan 6, 2025
I'm happy to post course materials for my class at UC Berkeley "Robots that Learn", taught with the outstanding assistance of @ToruO_O. Lecture videos at youtube.com/playlist?list=… Lecture notes & other course materials at robots-that-learn.github.io
6.7K
Toru
@ToruO_O
Sep 23, 2023
Update -- to appear in NeurIPS 2023 : D TL;DR - a simple, general, flexible framework for novelty-based RL exploration "MIMEx: Masked Input Modeling for Exploration" arxiv: arxiv.org/abs/2305.08932 code: github.com/ToruOwO/mimex w/ @ajabri [a rare two-author & student-only work]
Toru
@ToruO_O
May 17, 2023
New work! 🧵=> In RL (and perhaps in life), exploration is about maximizing one's knowledge of the environment so as to act more optimally in the long run. It is the *only* way through which an agent can learn to improve upon itself from scratch, and thus of central importance.
arxiv.org
MIMEx: Intrinsic Rewards from Masked Input Modeling
Exploring in environments with high-dimensional observations is hard. One promising approach for exploration is to use intrinsic rewards, which often boils down to estimating "novelty" of states,...
12K
Toru
@ToruO_O
Mar 21, 2025
Thanks for inviting -- it was great fun chatting with Mike & Chris! And lowkey announcement - the code for our last sim-to-real RL work has been released, feel free to check it out and see how it can help with your next project ;)))
RoboPapers
@RoboPapers
Mar 20, 2025
Ep3 with @ToruO_O on toruowo.github.io/recipe/ (Sim-to-Real Reinforcement Learning for Vision-Based Dexterous Manipulation on Humanoids). Co-hosted by @chris_j_paxton & @micoolcho
00:00
GitHub - ToruOwO/twisting-lids: Twisting Lids Off with Two Hands [CoRL 2024]
From github.com
4.9K
Toru
@ToruO_O
May 17, 2023
New work! 🧵=> In RL (and perhaps in life), exploration is about maximizing one's knowledge of the environment so as to act more optimally in the long run. It is the *only* way through which an agent can learn to improve upon itself from scratch, and thus of central importance.
18K
Toru
@ToruO_O
May 7, 2025
Really excited about the amazing speaker lineup! Come join our workshop on June 25th this year ✨
Youngwoon Lee
@YoungwoonLee
May 6, 2025
#RSS2025 Workshop on Whole-Body Control & Bimanual Manipulation: Applications in Humanoids & Beyond 🤖 Join us with our amazing lineup of speakers from academia & industry! Share your latest work by May 23 wcbm-workshop.github.io @RoboticsSciSys
4.5K
Toru
@ToruO_O
Feb 28, 2025
kind of funny that X flagged one of my humanoid videos as adult content... well I have to say it's a bare robot indeed🤣 @elonmusk ?
3.2K
Toru
@ToruO_O
Jul 31, 2023
Weekly research progress: didn't train any neural network but I proudly present this beautiful camera mount
6.4K