Yangcen Liu

44 posts

Yangcen Liu banner
Yangcen Liu

Yangcen Liu

@Randle_Liu

Incoming CS PhD @ Uchicago | Robo Master @ Georgia Tech | Robotics, Computer Vision

Atlanta, GA Katılım Kasım 2022
340 Takip Edilen145 Takipçiler
Sabitlenmiş Tweet
Yangcen Liu
Yangcen Liu@Randle_Liu·
What if one unified method helps robots learn from human videos across many tasks, many robots? Meet ImMimic: Cross-Domain Imitation from Human Videos via Mapping and Interpolation (CoRL 2025 Oral Presentation🏆) @ICatGT Check it here sites.google.com/view/immimic!
English
6
24
78
26.8K
Yangcen Liu retweetledi
Suning Huang
Suning Huang@suning_huang·
🤖Low-data post-training can teach a VLA policy a new robot skill. But it also makes it too attached to the training demos. We call this lock-in🔒: the policy can execute the post-training task, yet fails to respond to seemingly obvious prompt changes. DeLock preserves steerability using only the policy’s own pretrained knowledge. No extra supervision needed!🚀🚀🚀 #Robotics #AI #EmbodiedAI #VLA
English
4
44
174
28.5K
Yangcen Liu
Yangcen Liu@Randle_Liu·
This is really going to be the future. Zero human2robot gap (visual&action). Portable to collect data for complex tasks.
MANUS™@ManusMeta

How humans move their hands is still the best blueprint for teaching robots to do the same. @mimicrobotics, a Switzerland-based physical AI company with roots in @ETH Zurich, builds dexterous robotic hands and the manipulation models that drive them. To train those models at scale, they use MANUS data gloves to capture human hand motion, turning real-world demonstrations into the datasets their manipulation policies learn from. #robotics #datacollection #EmbodiedAI

English
0
0
4
204
Woo Chul Shin
Woo Chul Shin@woochulshin1726·
What if your robot could plan tasks it has never seen before without ever being retrained? Meet Compositional Visual Planning via Inference-Time Diffusion Scaling (ICLR 2026 🏆) comp-visual-planning.github.io If you are in Rio🇧🇷 visit us! Sat, 04/25/26 6:30-9:00 AM PDT Pavillion 4 #4203
English
2
7
38
8.4K
Yangcen Liu retweetledi
Project Aria @Meta
Project Aria @Meta@meta_aria·
⚡️EgoVerse is a first-of-its-kind, collaborative ecosystem for human-to-robot learning. The consortium leverages Project Aria to capture high-fidelity, egocentric human data — including 3D hand and head poses — to train next-gen robot manipulation policies. With over 1,300 hours of data across 2,000+ tasks, EgoVerse is a prime example of how the Aria Research Kit is being used by our partners to accelerate the future of embodied AI. Learn more: 🔗egoverse.ai 📰 arxiv.org/abs/2604.07607 Apply for the Aria Research Kit: projectaria.com/research-kit/ #MachineLearning #Robotics #ProjectAria #EgoVerse #ComputerVision @simar_kareer , @ryan_punamiya , @RogerQiu_42, @XiongyiCai , @yexelal
English
2
44
226
17.3K
Yangcen Liu
Yangcen Liu@Randle_Liu·
@teostealth There are more than 1300 hours of collected data. As for time to collect for all is not reported. Maybe double the time here?
English
0
0
0
17
Yangcen Liu
Yangcen Liu@Randle_Liu·
What an insightful essay! So amazing to work with Danfei for two years and explore at the frontier of learning from human behaviors. Zero-shot (bridge visual/action gap), co-design and VAM (H2R generation) three topics I am or will work on.
Danfei Xu@danfei_xu

x.com/i/article/2021…

English
1
2
21
4.6K
Yangcen Liu retweetledi
Haozhi Qi
Haozhi Qi@HaozhiQ·
Excited to join an amazing lineup as an invited speaker at the Manipulation Robustness workshop @ #ICRA2026 🚀 If you’re working on robust manipulation, consider submitting manipulation-robustness.github.io/icra2026/
Yifei Dong@YifeiDong314159

🚀 Excited to announce the Workshop on Manipulation Robustness at #ICRA2026 in Vienna 🇦🇹 (June 5)! 🌐 manipulation-robustness.github.io/icra2026/ Despite rapid progress, achieving robust, contact-rich manipulation in real-world uncertainty remains a core challenge. This workshop connects fundamental principles with practical pathways toward robust manipulation systems. 🎤 Invited speakers: @andrea_bajcsy, @HaozhiQ, @Jan_R_Peters, @SongShuran, Jie Song, Josie Hughes, Oliver Brock 📝 Call for Papers: Short papers / extended abstracts (2–4 pages) 📅 Deadline: May 5, 2026 🔗 Submission portal: OpenReview (openreview.net/group?id=IEEE.…) 🏆 Best Paper Awards (two robotic arms🤖 and cash prize) — sponsored by @ARXrobotics! Organized by @YifeiDong314159, @Kingchou007Li, @Hui_Zhang_eth, Teng Xue, @s_zhanyi, Quan Gao, Manuel Baum, @lujieyang98, Sylvain Calinon, @ftpokorny, @ChengXianyi. #Robotics #Manipulation #ICRA

English
0
2
25
3.1K
Yangcen Liu retweetledi
Fanqi Lin
Fanqi Lin@lfqirrrrr·
🧐 What actually works? Key takeaways from our single data modality analysis: (1) Co-training with diverse VL & cross-embodiment data substantially boosts generalization 🚀 to distribution shift, unseen tasks, and language following, while discrete action token variants offer no benefits. (2) Interestingly, the most effective co-training modalities are essentially diverse VL (see Fig. 👇), suggesting strengthening the VLM backbone directly translates to better policies. (3) Across all modalities, we observe no significant impact on in-distribution performance.
Fanqi Lin tweet media
English
1
2
11
3.7K
Yangcen Liu retweetledi
Justin Yu
Justin Yu@uynitsuj·
[Accepted to ICRA 2026!] 🚀 Introducing EgoMI: An egocentric manipulation interface that captures synchronized 6-DoF head and hand trajectories from egocentric human demonstrations! Transfers to IL policies zero-shot w/o visual augmentation or on-embodiment data. 1/n
English
6
26
245
31.4K
Yangcen Liu
Yangcen Liu@Randle_Liu·
Congrats to my friend and “mentor” Shuo! Another wonderful work from our lab for NeurIPS2025. Sim2Real and Human2Robot can be formulated as classic domain adaptation problems, and still a lot to explore in this line.
Shuo Cheng@ShuoCheng94

Can large-scale sim data enable real-world generalization?🤔 In our new work, we introduce a generalizable domain adaptation setting, where policies must handle real-world situations never presented in the real training data. (1/n)

English
0
0
0
246
Yangcen Liu
Yangcen Liu@Randle_Liu·
I'll be at NeurIPS next week in sunny San Diego. Would love to meet and chat—come say hi!
English
1
0
2
172