🧠Model-Based RL shows promises but has seen limited success in real-world robotics. 🌎Introducing Robotic World Model, a black-box end-to-end neural dynamics model that bridges this gap, where policies are trained purely in imagination. NeurIPS Conference 🎯sites.google.com/view/roboticwo…

Sim-to-real learning for humanoid robots is a full-stack problem. Today, Amazon FAR is releasing a full-stack solution: Holosoma. To accelerate research, we are open-sourcing a complete codebase covering multiple simulation backends, training, retargeting, and real-world

(1/5) Open Publication and the Prisoners dilemma. The prisoner’s dilemma game can offer us much insight into the publication strategy of leading industrial AI research labs. Suppose there are two labs A and B. If A and B both publish they are both better off than if they both

Wow!!

Here it is! Such a smooth policy! Also, I feel door opening makes an excellent case for a task challenging to achieve with teleop. Also, love the tagline “We are opening the sim-to-real door”

Very impressive! Pretty close to one of my old works: arxiv.org/abs/2208.13856 :)

Interesting!

🙇🙇

The toughest moment in a PhD: >spend a year building smth you’re proud of >travel across world with advisors’ support to share it >when your moment finally comes >your mic gets cut because the previous schedule ran late Heartbroken, but thanks to whom stayed for me NeurIPS Conference

Very insightful!

On their own, vision-language-action models are powerful tools for general robot skills that show impressive generalization. However, they don’t achieve useful levels of reliability on valuable manipulation tasks. Wenli Xiao 🏖️NeurIPS teaches us one way to achieve this reliability:

Generalist robots need a generalist evaluator. But how do you test safety without breaking things? 💥 🌎 Introducing our new work from Google DeepMind: Evaluating Gemini Robotics Policies in a Veo World Simulator veo-robotics.github.io 🧵👇

Wow!

Excited to share Veo Robotics! We illustrate how we can use Veo to construct a generative robot simulator that can simulate policy execution, predict policy performance and OOD generalization!

Crazy!

So good!

Mindblowing. Takeaway: ❌ Human data unlocks robotics ✅ Massive robot data unlocks the ability to use human data

If it works well enough, this unlocks internet data for robot training!