🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

People don’t realize the economic shock that’s about to happen over the next few years

Surgical Robot Transformer (SRT): Imitation Learning for Surgical Tasks Chelsea Finn Ji Woong Kim Johns Hopkins University Stanford University #roboticSurgery #MedEd #MEDX #Surgery #robotics surgical-robot-transformer.github.io

WOW 🤯 🤯 Da Vinci robot learning surgical manipulation tasks via imitation learning. The researchers successfully executed three fundamental surgical tasks, including tissue manipulation, needle handling, and knot-tying.

Excited to share that our paper Evaluation and mitigation of cognitive biases in medical language models was published in npj Digital Medicine In this work, we ask the question: can LLMs be easily fooled with simple cognitive bias-inducing prompts during patient diagnosis? To

The most insane gait I've seen on a humanoid. Walking with locked knees is much more energy efficient as the motors don't need to be engaged all the time. Public info of EngineAI: - team of 36 - raised ~14M USD - investment from SenseTime, Hefei province - founded Oct 2023

RL in the real world presents some big challenges, but also some really big opportunities. In our new work, HIL-SERL, Charles Xu, Jeffrey Wu, Jianlan Luo show that real-world RL can learn a huge range of precise and robust tasks, and perform them much faster than imitation.

Gary Guthart (CEO of Intuitive) and I propose the phrase “Augmented Dexterity” to describe systems where surgical subtasks are controlled by a robot under the close supervision of a human surgeon (vs. "supervised autonomy" which can be intimidating). Go Bears! UC Berkeley

excited to share what we've been up to soon!

At Conference on Robot Learning, I’m giving a talk on π₀ post-training, i.e. how we got the robot to fold laundry. Tomorrow/Saturday at 1:45 pm at the WCBM workshop: wcbm-workshop.github.io

Lots of memorable quotes from Jitendra MALIK at CoRL, the most significant one of course is: “I believe that Physical Intelligence is essential to AI” :) I did warn you Jitendra that out of context quotes are fair game. Some liberties taken wrt capitalization.

The first results from a π collaboration! 🤖☕🤖☕

One way to test the robustness of the policy is to mess with the robot. Some of the π₀ corrections are pretty impressive: Messing with the robot by throwing another shirt

Robot learning has been following in the footsteps of LLMs in many aspects, but one element stands apart - action chunking Why is it crucial? Any limitations? How to improve it? Introducing our paper Bidirectional Decoding: Improving Action Chunking via Closed-Loop Resampling🧵

A little bit of a history of the laundry folding task Physical Intelligence We started the company in March. By May we had our first relatively simple shirt folding policy. Starts from a simple state: t-shirt flat on the table Yes, it's Chelsea Finn proudly evaluating her policy 🦾

🚀🔬 Introducing Agent Laboratory: an assistant for automating machine learning research Agent Laboratory takes your research ideas and outputs a research paper and code repository, allowing you to allocate more effort toward ideation rather than low-level coding and writing 🧵

Simplify data collection with low cost,hand-held parallel jaw gripper #Pika AgileX Robotics Pika Gripper: ±1.5mm accuracy, dual-camera system Pika Sense: Lightweight (550g) Support ROS1/2 Perfect for researchers and developers driving innovation. #Robotics #EmbodiedAI #ROS

Excited to release FAST, our new robot action tokenizer! 🤖 Some highlights: - Simple autoregressive VLAs match diffusion VLA performance - Trains up to 5x faster - Works on all robot datasets we tested - First VLAs that work out-of-the-box in new environments! 🧵/