Wenlong Huang

@wenlong_huang

PhD Student @StanfordSVL. Previously @MIT_CSAIL @Berkeley_AI @GoogleDeepMind @NVIDIARobotics. Robotics, Foundation Models, Spatial Intelligence.

Stanford, CA

wenlonghuang.com

Replies Reposts Media

Pinned
Wenlong Huang
@wenlong_huang
Jan 8
What if we can simulate an *interactive 3D world*, from a single image, in the wild, in real time? Introducing PointWorld-1B: a large pre-trained 3D world model that predicts env dynamics given RGB-D capture and robot actions. 🌐 point-world.github.io from @Stanford @nvidia
00:00
31
Wenlong Huang
@wenlong_huang
Aug 26
@deepakpathak always has the best vision on how to make robots generalize, which I learned tremendously from during my years working with him (and continuously so!). And video prompting does feel like a key unlock towards in-the-wild robots. Huge congratulations to the team
Skild AI
@SkildAI
Aug 25
Introducing S1, our new foundation model that learns from one example. It can be taught 10-minute long tasks that it has never seen before, from one video prompt without any fine-tuning. Watch S1 operate in real-time via in-context learning:
00:00
3
Wenlong Huang
@wenlong_huang
Jul 24
What should be the actions when robots are learning from videos? Our answer: use 𝗮𝗰𝘁𝗶𝗼𝗻𝘀 𝗶𝗻 𝘁𝗵𝗲 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗼𝗳 𝘃𝗶𝘀𝗶𝗼𝗻. Introducing 𝗠𝗮𝘀𝗸𝗲𝗱 𝗩𝗶𝘀𝘂𝗮𝗹 𝗔𝗰𝘁𝗶𝗼𝗻𝘀 (𝗠𝗩𝗔), a technique to turn pre-trained video models into a robot world model
Hadi Alzayer
@HadiZayer
Jul 24
We trained a video world model on just 15 hours of video of a single-arm robot. 🧵 It generalizes zero-shot to unseen embodiments (and even orangutans). And it picked up something we never trained for: give it the object motion you want, and it synthesizes the robot motion that
00:00
1
Wenlong Huang
@wenlong_huang
Jul 21
Very exciting news! Congratulations @YunzhuLiYZ and @drfeifei !!
Fei-Fei Li
@drfeifei
Jul 21
The world is not just made of words, and spatial intelligence was never just about perceiving and generating worlds. It's about interacting with them. Today, SceniX is joining World Labs. 🌎🤖👇
00:00
Wenlong Huang
@wenlong_huang
Jul 16
Though not attending RSS in-person this year, I will give a remote talk and join the panel discussion at the Data-Centric Robotics Workshop today. I will talk about “Scaling Robotics with Counterfactuals”, a different take on how robot data should be scaled, and one that might
1