robot learning @ mit csail
Joined November 2019
- Tweets59
- Following186
- Followers558
- Likes144
Pinned Tweet
Every robot already carries contact sensors: its own joints. When a robot touches something, its joints drift slightly off their commands, and the encoders pick it up. Legged robots and humanoids already leverage this signal to walk blind - but is it enough for blind precise assembly?
We find that it is! Our policies can insert pegs, mesh gears, and thread nuts using only the robot's joint positions - no camera, no tactile, and no object poses. And the peg policy works zero-shot in the real world! 🧵
w/ my amazing co-first author @younghyo_park, and advised by @kushalk_ and @pulkitology.
I’ve been learning from Toru’s scientific writing for some years now, and it’s similarly valuable to read her thoughts on life :) highly recommend!
How should we approach research, when the very field we work on is going through rapid and potentially fundamental changes?
I recently gave a talk at the Stanford Vision Seminar, reflecting on my PhD journey and how AI is changing the way I think about research. I adapted part of it into a new blog post, organized around three questions:
1. What should we work on?
2. How should we work with AI?
3. What is a PhD for?
I hope some of my reflections are helpful. The talk ended with an open discussion, and I’d love to continue it here; curious how others are thinking about these questions.
Personally -- as much as there's FOMO, I also think life is very long, and we have time.
toruowo.github.io/blog/posts…
Antonia Bronars retweeted
We present OpenTac, an open, full-stack ecosystem for 3D tactile sensing on any robot surface.
Anyone can design and fabricate high-density tactile skins for any part of their robot, then read them out, simulate them, and share data across different sensors.
Antonia Bronars retweeted
Incredible stuff… doing precise manipulation is hard with force torque sensors.. this is a whole different level!!
Every robot already carries contact sensors: its own joints. When a robot touches something, its joints drift slightly off their commands, and the encoders pick it up. Legged robots and humanoids already leverage this signal to walk blind - but is it enough for blind precise assembly?
We find that it is! Our policies can insert pegs, mesh gears, and thread nuts using only the robot's joint positions - no camera, no tactile, and no object poses. And the peg policy works zero-shot in the real world! 🧵
w/ my amazing co-first author @younghyo_park, and advised by @kushalk_ and @pulkitology.
Cheers, Eka team! 🚀💥 behaviors are looking fast, reactive, and smoooooth
@EkaRobotics has raised a $150M Series A.
Generality v/s Performance is the fundamental tension at the frontier of robotics. Generality without performance is just a prototype. Performance without generality is just a single-purpose tool.
Customers shouldn’t have to choose. Eka breaks the tradeoff. Powered by our Vision-Force-Action (VFA) model, we deliver the performance and versatility customers need to put robots to work every day.
Eka is built on three convictions:
⚡️Force is the language of manipulation.
🚀 Reinforcement learning is how dexterous manipulation scales.
🎂 System 1—fast, reliable physical execution—is the bottleneck to solve.
With this funding, we will scale our model and accelerate customer deployments.
Grateful to be building alongside a dream team and investors who share our ambition.
Antonia Bronars retweeted
I really appreciate how hard it is to get this kind of thing working! 🥲
Big fan! Another wonderful paper from the Improbable fam
Every robot already carries contact sensors: its own joints. When a robot touches something, its joints drift slightly off their commands, and the encoders pick it up. Legged robots and humanoids already leverage this signal to walk blind - but is it enough for blind precise assembly?
We find that it is! Our policies can insert pegs, mesh gears, and thread nuts using only the robot's joint positions - no camera, no tactile, and no object poses. And the peg policy works zero-shot in the real world! 🧵
w/ my amazing co-first author @younghyo_park, and advised by @kushalk_ and @pulkitology.
Antonia Bronars retweeted
Can a robot assemble tight-clearance parts using only its proprioception? No camera, no tactile sensor, no F/T sensor, no joint torque sensing, no priors over the object geometries or poses: just joint position readings.
It actually can! Blind policies trained in simulation insert pegs, mesh gears, and thread nuts, and the peg policy works zero-shot in the real world. Here's what we found 🧵
Joint work with @BronarsToni, my wonderful co-first author, who's on the job market right now! Advised by @kushalk_ and @pulkitology.
9/ In short: joint sensing is a powerful modality for these fine-grained, precise assembly tasks. We show that a blind policy trained only in simulation is capable of searching intelligently through its own contacts to insert pegs in the real world. We show that gear meshing and nut threading can be learned similarly, and analyze the tradeoffs in robustness and speed compared with visual policies.
For more details, check out our project website and our paper!
🔗 website: younghyopark.me/blind-propri…
📄 arXiv: arxiv.org/abs/2610.23849
⚙️ mjfit: github.com/younghyopark/mjfi…
10/ On a personal note, this project is the final one of my PhD. It's been the most incredible privilege to work with folks at MIT and beyond over the past 6 years, and I can't wait for what's next! I'm on the job market for robotics research scientist positions starting in 2027 - so if you're hiring, I'd love to talk! tonibronars.github.io/
Nolan's done it again! Super cool work that tackles one of the most fundamental challenges in sim2real - accurate actuator modeling (even harder in the presence of contact).
(1/9) Same rewards. Less domain randomization. Higher jumps. Stable landings.
Just by improving the actuator model.
Introducing Contact-UAN: a residual world model for humanoids, learned from proprioception. No torque sensing or motion capture.
Project website: contact-uan.csail.mit.edu.
Antonia Bronars retweeted
As humanoid robots become part of our everyday lives, they need to do more than just perform impressive motions. They should be safe, fun to interact with, and respond naturally to unexpected contact, without sacrificing performance.
Our latest work, Compliant GMT (C-GMT), brings whole-body compliant responses to general motion tracking.
🌐 More details: venki1899.github.io/c-gmt/
Work done at MIT Improbable AI Lab with @michl_wang, Jin, @pulkitology.
Antonia Bronars retweeted
🤖 Why do generative policies beat regression policies? Multi-modality? Manifold adherence? Or more?
Our analysis suggests: highly imbalanced gradients under MSE!
A loss change closes the gap: no denoising, competitive performance, faster inference.
📄 arxiv.org/abs/2610.12231
Antonia Bronars retweeted
There’s lots of buzz around agentic harnesses for robots, particularly for long horizon tasks that require complex reasoning and memory. But what will it really take to turn a reasoning agent into a reactive, reliable and low-latency robot policy?
In our new paper, workspace models, we design a new memory architecture that acts as a latent harness for stronger reasoning models. Here’s why we think this might be the way forward:
(1/9)
