Leonardo Perez

math @pucp | robot learning

[Leonardo Perez @leoperzz]

Amazing experience that we had at #ICRA2026. My favorite photo

[Aryan Mangla @Aryan_Mangla_]

We won 1st place in Logistics Picking track at the #ICRA2026 Vienna Site of What Bimanuals Can Do 2026 @WBCDCompetition

It focused on whole-body humanoid logistics picking task

The journey and experience was just amazing!

@leoperzz @autobrik @raulb4s @mbrq_13 @ubillus83797

[Leonardo Perez @leoperzz]

@aryanmadhaverma looks really good! 👀

[Artificio @Artificio_Org]

Debugging

[Leonardo Perez @leoperzz]

@iyanmoonyang I think vision is mainly used for trajectory planning. Your brain can still adjust your movements using only sensory feedback from the ground

[Leonardo Perez @leoperzz]

@nanliuuu yes, in fact, pi07 only predicts the next state as an image https://t.co/GFBlHiZm1I

[Lester Li @sizhe_lester_li]

Robot learning is moving beyond policies built for one robot, one scene, one task.

At MIT, we’re exploring a different path: turning video world models into embodiment-agnostic robot policies.

Introducing VERA: a 14B video-to-action system that controls robots across embodiments, skills, and environments.

From zero-shot pick-and-place on a real Panda arm to contact-rich cube reorientation with a 16-DoF robotic hand.

Different robots. Different environments. Different tasks.
Same video planner. Same weights.

We’re open-sourcing everything so you can fine-tune VERA for your own robot setup too. Deep dive in the thread:

🔗 https://t.co/hzuYZ2m5lS
🧵 (1/7)

[SAGE @SAGE_1125]

We recently wrote a short blog on the mathematical essence behind three common World Model paradigms in Robot Learning.

It looks at Future-conditioned / IDM-style, Single-backbone, and MoT-style models from the lens of probabilistic modeling and structured optimization.

[Leonardo Perez @leoperzz]

@iyanmoonyang https://t.co/HWYnt2yNqa this is really good

[Leonardo Perez @leoperzz]

@iyanmoonyang https://t.co/fFlMQC4JrF :)

[T.Yamazaki @ZappyZappy7]

『ロボットに「人間の動きを真似させる」だけでなく、ロボット自身の体格・重さ・関節・バランスに合わせて、現実に動ける形へ変換するための技術』
押す・蹴る・運ぶ全身作業の学習を加速
https://t.co/BhDvmLmhqy

#HumanoidRobot #PhysicalAI #RobotLearning #EmbodiedAI #DynaRetarget #ATARI_LAB

[Yunsong Zhou @Yunsong_Zhou]

Excited to share that SIM1 has been accepted to ECCV 2026! 🎉
Huge thanks to our amazing team for making this possible.
See you in Sweden! 🇸🇪 #ECCV2026

[Ethan Clark @ethanmclark1]

Working in robotics right now is what I imagine working with language models felt like in 2023. Everyone throwing things at the wall to see what sticks

Pixel prediction (Cosmos), action prediction (VLA), reward prediction (TD-MPC), and representation prediction (JEPA). Different paths for the same problem

The recipe that won in language was self-supervised pretraining at internet scale then light finetune on top. Only representation prediction runs that playbook. It learns from action-free video data so you can pretrain on YouTube and egocentric data then add a control layer. Everything else needs action-labeled data that doesn't scale

As an RL maximalist, I used to hate LeCun's cake. Turns out he was right all along which is how I ended up a JEPA truther

[Seungjae (Jay) LEE @JayLEE_0301]

Can a model trained purely on video — with zero action labels — match VLAs trained on massive action-labeled datasets?

Meet µ0 (Mew-Zero): a world model that learns a "physical language" for robots.

Here's why we're excited 🧵

[NVIDIA Robotics @NVIDIARobotics]

@Aryan_Mangla_ @WBCDCompetition @leoperzz @autobrik @raulb4s @mbrq_13 @ubillus83797 @0xnonhuman @XYZRoboticsInc @NVIDIAAI @DrJimFan Congrats!

[Leonardo Perez @leoperzz]

@NVIDIARobotics @Aryan_Mangla_ @WBCDCompetition @autobrik @raulb4s @mbrq_13 @ubillus83797 @0xnonhuman @XYZRoboticsInc @NVIDIAAI @DrJimFan Thank you for SONIC. It's amazing🫡

[Leonardo Perez @leoperzz]

Is it possible to use the same model to do this and do laundry, for example? One of the main problems, I think, is how we can achieve really high-frequency policies

[Space and Technology @spaceandtech_]

Researchers from The University of Hong Kong and Kinetix AI have developed a humanoid robot system called SMASH that can play real table tennis using only onboard cameras.

The robot tracks the ball in real time without using external cameras or motion-capture systems.

It can perform powerful smashes, quick side movements, and low crouching saves using full-body coordination.

[Lucky Robots @luckyrobots]

We are super excited to share with you our initial release of Lucky Engine. We are building a robotics engine from the ground up to be what we wished we could find in a simulator before

[Jason Liu @JasonJZLiu]

💥Introducing FACTR 2, learning external force sensing on commodity robot arms without needing dedicated sensors.

We show that learned force signals enable force-feedback teleop on low-cost arms and improve BC policies.

FACTR 2 consists of:
1. Neural External Torque (NEXT): learns external forces without needing dedicated force sensors.
2. Force-Informed Re-Sampling Training (FIRST): uses the learned force signal to identify task-critical regions and upsample them during training.

w/ @StevenOh_ @_tonytao_

🧵(1/N)

[Leonardo Perez @leoperzz]

The most impressive demo I’ve seen at ICRA! Congrats guys and thank you for the socks too

[Flexion Robotics @FlexionRobotics]

We ran 300 fully autonomous live demonstrations over 3 days at ICRA 2026.

The task: a humanoid navigating stairs, picking up a box from the floor and placing it on a table. Simple to describe, but hard to execute reliably when your robot is making every decision on its own at a conference with new surroundings and a crowd watching live.

This is just a glimpse. We've been pushing our stack much further and we'll be sharing more very soon.

More information in the thread.

#HumanoidRobots #ICRA2026 #Flexion

[Junzhe (JJ) He @JayHe748646]

Excited to share our recent work on whole-body humanoid locomotion for challenging terrain traversal!

Diffusion-based planner + RL WBC = general purpose locomotion controller

Led by @ctki49 @mxu_cg @KehanWen170077 at @leggedrobotics and @xbpeng4.