Excited to share my recent work with @gabe_mrgl , Martin Pettico, and @pulkitology . We’re pushing the limits of whole-body control to make robots faster, stronger, and more athletic!
🤖 We introduce Ambient Diffusion Policy, a simple and principled method for training policies with suboptimal data in robotics.
Suboptimal data is everywhere in robotics…
❌ Data filtering is wasteful
❌ Co-training learns both good and bad features
✅ Ambient Diffusion Policy selectively learns useful features via noise-dependent data usage
👇🧵(1/5)
It was a great pleasure to give a guest lecture, "Sim2Real Robot Learning: A Holistic Overview," for @chelseabfinn's CS224R at @Stanford. The lecture covered the motivation, history, key methods, and future outlook of sim-to-real robot learning. Slides: https://t.co/NZRk1WvQOX
Eka means unity -- “one,” in Sanskrit and “first” in Finnish.
We’re building intelligence for the physical world in its native language: forces.
Until now, robotics faced a tradeoff — generality or speed. The real world requires both. Robotics also faced a data problem.
Our Vision–Force–Action (VFA) model — the first of its kind — breaks the generality-speed tradeoff and the data barrier.
It's a new foundation uniting performance, generality, and safety for putting capable robots in everyone's hands.
Today, I am excited to share our journey of pushing robots beyond human limits.
Today, dexterity becomes scalable.
Today, I welcome you to the Era of Eka.
Co-founded with @haarnoja, and so thrilled and grateful to be working with a dream team at @EkaRobotics.
Learn more: https://t.co/QYQ6x2Etyi
What's different between these two BC policies? It's the same architecture, training budget, and data collection setup — the only difference is the controller gains!
Controller gains are an understudied design parameter in robot learning. In our new work (w/ @BronarsToni*, @pulkitology), we show how they act as an inductive bias across BC, RL, and Sim2Real transfer, with real consequences on performance. Here's what we found 🧵
* Equal Contribution
📄arxiv: https://t.co/SMYgh7i8cA
🔗website: https://t.co/cLCd1FYCdJ
Compliance is critical not only for safety but also human-like object interaction. Hope to see more of the sim2real RL community explore compliant control. Great working with @gabe_mrgl, @michl_wang, and @pulkitology.
Excited to share that I'll be joining @UTAustin in Fall 2026 as an Assistant Professor with @utmechengr@texas_robotics!
I'm looking for PhD students interested in humanoids, dexterous manipulation, tactile sensing, and robot learning in general -- consider applying this cycle!
For agents to improve over time, they can’t afford to forget what they’ve already mastered.
We found that supervised fine-tuning forgets more than RL when training on a new task!
Want to find out why? 👇
What’s keeping robot arms from working like human arms?
They're big, slow, have the wrong joints, and can't conform to their environment. DexWrist solves all of these issues and simplifies learning constrained, dynamic manipulation👉 https://t.co/qxNJVVx5hV
The next frontier for AI shouldn’t just be generally helpful.
It should be helpful for you!
Our new paper shows how to personalize LLMs — efficiently, scalably, and without retraining.
Meet PReF (https://t.co/XaLZAwimse)
1\n
Excited to share my recent work with @gabe_mrgl , Martin Pettico, and @pulkitology . We’re pushing the limits of whole-body control to make robots faster, stronger, and more athletic!
@AayushDesh@gabe_mrgl@pulkitology Thanks! I've been working on whole-body control for about a year and a half now. Project timelines can vary a lot though depending on experience level and how much past work you can leverage
@albericlajarte@gabe_mrgl@pulkitology Thanks! -- It takes in a history of past (simulated) position errors and velocity measurements. We use the hardware data for the reward function and to set the initial state of the sim during UAN training.
Presenting Unsupervised Actuator Nets (UANs) that push the limits of agile whole-body control without the need for reward shaping!
⚡️ UANs reduce the sim2real gap in robot's motors removing the need for reward design to bridge the sim2real gap.
⚡️ A pre-trained whole-body controller uses reference motion as a hint to maximize task performance!
Learn more: https://t.co/f68scKkU8Z
Work led by @nolan_fey , @gabe_mrgl in collaboration with Martin.
To address exploration challenges, we fine-tune a foundational whole-body controller with simple task rewards, using reference trajectories as helpful hints.