Long Le

@LongLeRobot

Student Researcher @GoogleDeepMind | PhD student @Penn. Working on robot learning.

Philadelphia, USA

vlongle.github.io

Replies Reposts Media

Pinned
Long Le
@LongLeRobot
Mar 13
Check out my most recent project that I've been working on for more than half a year -- co-led with Tim @TimSong52005757! We showed a flexible and effective recipe for improving generalist VLAs leveraging non-robotic foundation models! The key is guidance in 3D! More details @
Kostas Daniilidis
@KostasPenn
Mar 13
VLA policies learn generalist robot behaviors from massive teleoperation datasets, hoping that the right behavior emerges. But they rarely use perception during training or inference: powerful foundation models of 3D geometry, semantics, or human motion are ignored.
00:00
Long Le
@LongLeRobot
Sep 4
Super thrilled that our paper has just been accepted to #CoRL2026! 🎉🥳 See you in Texas!
Kostas Daniilidis
@KostasPenn
Mar 13
VLA policies learn generalist robot behaviors from massive teleoperation datasets, hoping that the right behavior emerges. But they rarely use perception during training or inference: powerful foundation models of 3D geometry, semantics, or human motion are ignored.
00:00
1
Long Le
@LongLeRobot
Jun 6
From PIXIE → UniPixie: physics from pixels becomes *generative*. UniPixie learns a controllable soft↔stiff spectrum of 3D physical properties from visual input, across simulators: MPM/LBS/spring-mass. #CVPR2026 Highlight ✨ Poster: ExHall F, 11:45 AM today. Come say hi!
1
Long Le
@LongLeRobot
May 9
Personal news: I’m joining @GoogleDeepMind NYC this summer as a Student Researcher on the Robotics team! 4 years ago, I left Google to start my PhD. Now I get to come back and work on humanoid robot learning. Full loop closure, as the SLAM nerds would say. NYC folks—say hi!
47
Long Le
@LongLeRobot
Mar 14
For this project, we’ve also optimized the speed significantly and shown real-time demos to several visitors coming to Penn including @chris_j_paxton and @DJiafei. Here’s a video of the robot finishing a task in 20 secs autonomously
00:00
Kostas Daniilidis
@KostasPenn
Mar 13
00:36
VLA policies learn generalist robot behaviors from massive teleoperation datasets, hoping that the right behavior emerges. But they rarely use perception during training or inference: powerful foundation models of 3D geometry, semantics, or human motion are ignored.
2