Tianming Wei

PhD student @Tsinghua_Uni | Prev. @SJTU1896. Interested in Robotics, Sim2real.

[Yufeng Tian @TianRainWind]

Eyes see. Hands feel. Many visuotactile methods rely only on concatenation, leaving room to better exploit alignment and complementarity for manipulation.
Excited to share our #ICRA2026 paper: "ViTaS: Visual Tactile Soft Fusion Contrastive Learning for Visuomotor Learning".

[Juncheng Mu @Juncheng_Mu]

Can robots learn a bimanual dexterous skill from just one video?
DexImit makes it possible 🏄‍♂️ — automatically converting human videos into physically plausible training data for bimanual manipulation, either from text2video models or real-world recordings.

[RoboPapers @RoboPapers]

Just collecting manipulation data isn’t enough for robots - they need to be able to move around in the world, which has a whole different set of challenges from pure manipulation. And bringing navigation and manipulation together in a single framework is even more challenging.

Enter HERMES, from @fancy_yzc and @still_wtm. This is a four-stage process in which human videos are used to set up an RL sim-to-real training pipeline in order to overcome differences between robot and human kinematics, and used together with a navigation foundation model to move around in a variety of environments.

To learn more, join us as @fancy_yzc and @still_wtm tell us about how they built their system to perform mobile dexterous manipulation from human videos in a variety of environments.

Watch Episode #45 of RoboPapers today, hosted by @micoolcho and @chris_j_paxton!

[Kun Lei @kunlei15]

Introducing RL-100: Performant Robotic Manipulation with Real-World Reinforcement Learning. https://t.co/FUqEQ26mzi

7 real robot tasks, 900/900 successes. Up to 250 consecutive trials in one task, running 2 hours nonstop without failure.

High success rate against physical disturbances, zero-shot, and few-shot adaptation

Our first step toward a deployable robot learning system.

[Zilin @zilin_kang]

🚀 Thrilled to introduce ERA: Entropy Regularizing Activation!
A new paradigm to control model entropy via output activations.
Delivering major performance boosts in:
🤖 LLM Reasoning
🤸 Continuous Control
🖼️ Image Classification

Thread below 🧵

[Kun Lei @kunlei15]

Very impressive！Focused sim2real mobile dexterous manipulation，human data is leveraged for the guidance in reinforcement learning.

[Tianming Wei @still_wtm]

In weekly meeting, our advisor Huazhe kept asking: "What's the real point of simulation for manipulation?" HERMES offers a potential answer. Our framework converts diverse human motion data into real robot behaviors via sim training, all without task-specific reward designing.

[Zhecheng Yuan @fancy_yzc]

👐How can we leverage multi-source human motion data, transform it into robot-feasible behaviors, and deploy it across diverse scenarios?

👤🤖Introduce 𝐇𝐄𝐑𝐌𝐄𝐒: a versatile human-to-robot embodied learning framework tailored for mobile bimanual dexterous manipulation.

[Yuanchen Ju @ju_yuanchen]

🧩#CVPR2025🌷Introducing Two By Two✌️: The First Large-Scale Daily Pairwise Assembly Dataset with SE(3)-Equivariant Pose Estimation.
🤖2BY2 helps robots master daily 3D assembly tasks—like plugging sockets or arranging flowers—across diverse objects!
🐨Co-lead by @yuqi_Beijing

[Yuanchen Ju @ju_yuanchen]

🍌We present DenseMatcher！
🤖️DenseMatcher enables robots to acquire generalizable skills across diverse object categories by only seeing one demo, by finding correspondences between 3D objects even with different types, shapes, and appearances.

[Galaxea Dynamics @GalaxeaDynamics]

Having fun with Galaxea AI’s A1 robotic arms! Experience smooth teleportation and unmatched precision. Watch how it combines dynamic control and strength to tackle any task effortlessly. #GalaxeaAI #Robotics #AI

[SourceRobotics @SourceRobotics]

Controlling a CM6 robot with a LEAP motion controller. J2 of the robot is a bit underpowered but we mainly tested our algorithm here that we will implement on our other robots. Data is shared at 125Hz between robot and Leap over local network.

[Huazhe Harry Xu @HarryXu12]

Drop the camera, and bang -- the robots do the job without camera calibration or additional data collection. We want robots to manipulate anywhere! Our new work, **Maniwhere**, enables robots to manipulate objects from any camera view with any background. Isn't this the generalization every roboticist is looking for? How do we achieve this with sim2real? You can check this thread https://t.co/WtFHQaXLwL from @fancy_yzc ! And the website is here: https://t.co/htrD0Suk3A

[Tianming Wei @still_wtm]

🥳Glad to share our new work **Maniwhere** !
Maniwhere enables your robot to generalize across diverse visual conditions and camera viewpoints in a zero-shot manner.🤖 No more worrying about your lab mates messing with the camera setup you've carefully designed ! 🤣

[Zhecheng Yuan @fancy_yzc]

How to make your robot handle diverse visual scenarios?🤔
Introduce our recent work: Learning to 𝑴𝒂𝒏𝒊pulate Any𝒘𝒉𝒆𝒓𝒆: 𝐌𝐚𝐧𝐢𝐰𝐡𝐞𝐫𝐞. It enables your robot to manage multiple visual disturbance types 🌈and step out of simulation to achieve sim2real transfer.🤖