Yuntao Ma (@yuntao144) 's Twitter Profile
Yuntao Ma

@yuntao144

Legged mobile manipulation

ID: 1840275038341099520

calendar_today29-09-2024 06:19:21

7 Tweet

137 Followers

79 Following

Yuntao Ma (@yuntao144) 's Twitter Profile Photo

The control policy was actually trained with a different arm + different gripper. It's amazing to see it working with the new setup😍

Science Robotics (@scirobotics) 's Twitter Profile Photo

A legged robot can play badminton #autonomously with humans with the help of a #ReinforcementLearning-based controller that coordinates perception and movements. Robotic Systems Lab Learn more in Science #Robotics: scim.ag/3Z1AYF7

Yuntao Ma (@yuntao144) 's Twitter Profile Photo

Table 4 is my personal favorite table across all the papers I’ve read in the past years. Super clean and convincing🥳

Yuntao Ma (@yuntao144) 's Twitter Profile Photo

Proud to share our latest work from Robotic Systems Lab ! A legged robot that rallies with humans in badminton 🏸📷 Whole-body agility, onboard perception, RL-trained.

Science Robotics (@scirobotics) 's Twitter Profile Photo

Scientists have developed a control framework that enables #Raibo, a legged robot, to perform #parkour-like maneuvers and navigate complex environments autonomously. KAISTPR Jemin hwangbo Read more in Science #Robotics: scim.ag/4dw3tAE

Yunlong Song (@realyunlong) 's Twitter Profile Photo

Swarm navigation at 20 m/s with no communication, no state estimation, & on a $21 computer? This paper combines deep learning + differentiable sim for zero-shot sim-to-real flight. It is an underrated breakthrough. Paper: arxiv.org/abs/2407.10648 #naturemachineintelligence

Yuntao Ma (@yuntao144) 's Twitter Profile Photo

It only snowed once before I left for vacation last winter and it stayed just two days. Perfect timing for our experiments! Among all aspects of this project, timing experiments are by far my favorite. Robotic Systems Lab #LeggedRobotics #RoboticManipulation #SnowyTesting

Andrej Karpathy (@karpathy) 's Twitter Profile Photo

+1 for "context engineering" over "prompt engineering". People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window

Takahiro Miki (@ki_ki_ki1) 's Twitter Profile Photo

We're organizing a workshop at humanoids about RL and model based control. Please submit an extended abstract for the poster session!!