Josh Citron

ME + CS @Stanford | Working in REALab

[Josh Citron @josh__citron]

Teleop systems are usually designed for a single embodiment, but do they have to be?

Introducing ModPack 🎒: a modular teleoperation interface for bimanual mobile robots.

A wearable backpack provides shared infrastructure, robot-specific leader arms adapt to different embodiments, and plug-and-play modules add capabilities like haptic feedback, active perception, and mobile manipulation. 🧵(1/n)

[Huy Ha @haqhuy]

Transformer Transformer will be presented at #CoRL2026 🚀 See y'all in Austin!

Website: https://t.co/s7eimKcHbb
Video: https://t.co/YGQXUSZWLY
Paper: https://t.co/Biyu0r1J7N

[Maximilian Du @du_maximilian]

When a robot learns a new behavior, what keeps it from forgetting the ones it already knows?

Prior work, especially on VLAs, shows that rehearsing past experiences can enable strong continual learning. But why does rehearsal work so well, and when does it fail?

We find that a small but ✨special subset✨ of rehearsed data largely determines whether old behaviors are retained or forgotten. We call these examples Memory Anchors. ⚓

Withholding just the top 10% of Memory Anchors from rehearsal increases forgetting by up to 4.5x. Increasing their presence, meanwhile, reduces task forgetting and enables a real robot to learn challenging task sequences.

Website: https://t.co/JhGATqmQ6F

Paper: https://t.co/i3PPO2NxP3

Curious? Read on! 🧵👇 (1/9)

[Josh Citron @josh__citron]

@Haoyu_Xiong_ Thanks Haoyu!

[Josh Citron @josh__citron]

Teleop systems are usually designed for a single embodiment, but do they have to be?

Introducing ModPack 🎒: a modular teleoperation interface for bimanual mobile robots.

A wearable backpack provides shared infrastructure, robot-specific leader arms adapt to different embodiments, and plug-and-play modules add capabilities like haptic feedback, active perception, and mobile manipulation. 🧵(1/n)

[Josh Citron @josh__citron]

The code and CAD files are fully open source along with a bill of materials for anyone to build. Thank you to my wonderful collaborators @reneezbizzz, @Liu_Zeyi_, and of course @SongShuran.

📄 Paper: https://t.co/gZFxRrjN0F
🌐 Website: https://t.co/hlr2EfJqQz
💻 Code: https://t.co/vsFaPgiL2O
📚 Documentation: https://t.co/fnCMYC7zNL

[Josh Citron @josh__citron]

Beyond policy training, ModPack supports open-ended teleoperation in everyday environments. We demonstrate three additional tasks across both robots: cooking food in a kitchen, recycling cans, and calling an elevator along with the previous tasks.

Highlights below, but you can find full demo runs on the project page in the teleoperation section.

[Xiaomeng Xu @XiaomengXu11]

Can we learn whole-body mobile manipulation directly from human demonstrations?

Introducing Whole-Body Mobile Manipulation Interface (HoMMI)

Egocentric + UMI, 0 teleop -> bimanual & whole-body manipulation, long-horizon navigation, active perception

https://t.co/CcZ9ZwfuFr

[Danfei Xu @danfei_xu]

Great working with you and the Stanford collaborators. Thanks for believing in the vision. Excited to keep building EgoVerse together!

[Josh Citron @josh__citron]

I am proud to share this new work with everyone, really happy with how it came together!

[Danfei Xu @danfei_xu]

Introducing EgoVerse: an ecosystem for robot learning from egocentric human data.

Built and tested by 4 research labs + 3 industry partners, EgoVerse enables both science and scaling

1300+ hrs, 240 scenes, 2000+ tasks, and growing

Dataset design, findings, and ecosystem 🧵