Yitang Li (@li_yitang) / X

Yitang Li

216 posts

Yitang Li

@li_yitang

Undergraduate in IIIS (Yao Class) at THU | Robot Learning; Motion synthesis | Applying for PhD-26Fall

Joined February 2025

Pinned
Yitang Li
@li_yitang
Jan 28
🎉BFM-Zero is accepted to ICLR 2026! 🌟In our framework, data guides exploration rather than constraining it. The robot is allowed to fall, roll, and freely explore anywhere it can reach. In unsupervised RL, no motion is inherently GOOD or BAD, and no motion should be directly
Yitang Li
@li_yitang
Nov 7, 2025
Meet BFM-Zero: A Promptable Humanoid Behavioral Foundation Model w/ Unsupervised RL👉 lecar-lab.github.io/BFM-Zero/ 🧩ONE latent space for ALL tasks ⚡Zero-shot goal reaching, tracking, and reward optimization (any reward at test time), from ONE policy 🤖Natural recovery & transition
00:00
32K
Yitang Li
@li_yitang
Nov 10, 2025
Check out Safe & Natural & Robust Recovery of BFM-Zero (highlight moments👇)🤣🤣🤣
00:00
00:38
Yitang Li
@li_yitang
Nov 7, 2025
Replying to @li_yitang
🌿 Smooth & well-regulated space enables: 1️⃣ Natural recovery under disturbance (it even runs a few steps🦿to regain balance when pushed🫱!) 2️⃣ Search-based few-shot adaptation 3️⃣ Meaningful latent-space interpolation See more details in our website~
146K
Yitang Li
@li_yitang
Nov 7, 2025
Meet BFM-Zero: A Promptable Humanoid Behavioral Foundation Model w/ Unsupervised RL👉 lecar-lab.github.io/BFM-Zero/ 🧩ONE latent space for ALL tasks ⚡Zero-shot goal reaching, tracking, and reward optimization (any reward at test time), from ONE policy 🤖Natural recovery & transition
00:00
81K
Yitang Li
@li_yitang
Jun 2, 2025
🤖Can a humanoid robot carry a full cup of beer without spilling while walking 🍺? Hold My Beer ! Introducing Hold My Beer🍺: Learning Gentle Humanoid Locomotion and End-Effector Stabilization Control Project: lecar-lab.github.io/SoFTA/ See more details below👇
00:00
61K
Yitang Li
@li_yitang
Oct 9, 2025
When I started research, I was amazed by DeepMimic—watching a simulated puppet learn to spin-kick😍After a year of exploration, that wonder made me believe even more in the power of motion priors and the human data intelligence. Now, that original magic is a toolkit for all😋
Jason Peng
@xbpeng4
Oct 8, 2025
Implementing motion imitation methods involves lots of nuisances. Not many codebases get all the details right. So, we're excited to release MimicKit! github.com/xbpeng/MimicKit A framework with high quality implementations of our methods: DeepMimic, AMP, ASE, ADD, and more to come!
00:00
7.8K
Yitang Li
@li_yitang
Sep 29, 2025
Finally done with the HoldMyBeer present task hahaha Big thanks to everyone who came — so fun seeing so many people crack up at Hold My Beer! hope you like my “busy” poster style😋 and also big thanks for all the support and photos!
5.3K
Yitang Li
@li_yitang
Jun 10, 2025
🎉 On my way to Nashville @CVPR ! ✈️Sooo--o excited to see my old friends, make new ones, and finally meet my online pals in person! 🥰 Feel free to drop by our poster anytime to chat or just wave hello👋! 🎤 Poster Session 🕔 5:00 – 7:00 PM 📍 Exhibit Hall D, #144 See more👇
00:19
Yitang Li
@li_yitang
Apr 18, 2025
Introduce our \CVPR 2025/ Learning Physics-Based Full-Body Human Reaching and Grasping from Brief Walking References 🌟 Ever wondered how to create lifelike human motions without mountains of Mocap data? 💡 Task-driven generated data + Limited(walking only!!!) realistic Mocap
7.5K
Yitang Li
@li_yitang
Apr 18, 2025
Introduce our \CVPR 2025/ Learning Physics-Based Full-Body Human Reaching and Grasping from Brief Walking References 🌟 Ever wondered how to create lifelike human motions without mountains of Mocap data? 💡 Task-driven generated data + Limited(walking only!!!) realistic Mocap
00:00
11K
Yitang Li
@li_yitang
Sep 26, 2025
Came to Seoul with Hold My Beer🍺! Come hang out and chat with me anytime😋 Spotlight 3 & Poster 2
Yitang Li
@li_yitang
Jun 2, 2025
🤖Can a humanoid robot carry a full cup of beer without spilling while walking 🍺? Hold My Beer ! Introducing Hold My Beer🍺: Learning Gentle Humanoid Locomotion and End-Effector Stabilization Control Project: lecar-lab.github.io/SoFTA/ See more details below👇
00:00
8.8K
Yitang Li
@li_yitang
Nov 7, 2025
Replying to @li_yitang
🌿 Smooth & well-regulated space enables: 1️⃣ Natural recovery under disturbance (it even runs a few steps🦿to regain balance when pushed🫱!) 2️⃣ Search-based few-shot adaptation 3️⃣ Meaningful latent-space interpolation See more details in our website~
00:00
144K
Yitang Li
@li_yitang
Nov 7, 2025
I always knew 👉representation matters — but only when I touched the robot👀, felt its responses, did I sense its magic. Most precious memories @TongheZhang01 @JimmyDai218776 they were there at the most critical moments, making it all possible!
Tonghe Zhang
@TongheZhang01
Nov 7, 2025
Was deeply involved in all the real world experiments that came into the paper, it's really amazing to see BFM-zero creating a truly generalist humanoid robot, that downloads motion skills and composite them like in the movie Matrix. Huge thanks to the teammates!!
9K
Yitang Li
@li_yitang
Sep 23, 2025
And it’s the most robust policy I’ve ever seen — a 35-minute nonstop demo of door traversal, even under terrain disturbances! 🔥
Haoyang Weng
@ElijahGalahad
Sep 22, 2025
Replying to @ElijahGalahad
HDMI is capable of: ✅ 67 consecutive door runs (34 minutes!) ✅ Box loco-manipulation with whole-body coordination ✅ Complex multi-stage behaviors like Truman’s Bow ⚡... all with a single set of rewards and observations!
00:00
6K
Yitang Li
@li_yitang
Oct 8, 2025
Absolutely love how UMI-on-Air brings the UMI to the sky ☁️🤖 😍 Using gradient feedback to make diffusion policies embodiment-aware — 🤩to see embodiment-agnostic learning become truly practical through guided diffusion Congrats @hgupt3 @Xiaofeng2Guo 🥰
Harsh Gupta 🇺🇸
@hgupt3
Oct 8, 2025
✈️🤖 What if an embodiment-agnostic visuomotor policy could adapt to diverse robot embodiments at inference with no fine-tuning? Introducing UMI-on-Air, a framework that brings embodiment-aware guidance to diffusion policies for precise, contact-rich aerial manipulation.
00:00
5.6K
Yitang Li
@li_yitang
Sep 21, 2025
Never imagined we can have this robust and expressive in ONE policy😳😳😳 Push, pull, payload, uneven terrain… no worries falling down!!!!!! Ok😄😄Time to chill and just play with this SAFE humanoid 🤖✨ Congrats!!! @Zhikai273
Zhikai Zhang
@Zhikai273
Sep 19, 2025
🚀 Looking forward to a foundtional humanoid motion tracker? 📢Meet Any2Track — Track Any Motions under Any Disturbances! 💡Any2Track shows SOTA performance on diverse motions under multiple disturbances. Project: zzk273.github.io/Any2Track/ Codebase: github.com/GalaxyGeneralR…
00:00
6.2K