Come find us at the RSS Workshop on Lifelong Robot Learning to learn more about how we can learn robotic skills directly from a long, unlabelled human video with zero model training or fine tuning by leveraging existing Foundation Models!

Yesterday at #RSS2024, Norman Di Palo presented Keypoint Action Tokens (KAT) … And today (right now), he’s presenting our new work which shows how robots can use KAT to learn from humans via *retrieval* and *execution*. Come to the Workshop on Lifelong Robot Learning for more!

When I was organising #CoRL 2021, we introduced demos for the first time (including my lab's own demo, below). I love that this then become a CoRL tradition, and I encourage everyone to think about what you could show this year (see tweet below). More manipulation demos please!

The future of robotics is in your hands. Literally. Excited to announce our new paper,✨R+X✨. A person records everyday activities while wearing a camera.A robot passively learns those skills. No labels, no training. Here's how. 👇

Excited to introduce Diffusion Augmented Agents (DAAGs)✨. We give an agent control of a diffusion model, so it can create its own *synthetic experience*.🪄 The result is a lifelong agent that can learn new reward detectors and policies, much more efficiently. Here's how. 👇

What’s next for LLMs + Robotics? Come and find out at CoRL this year, where we’re organising the 3rd Workshop on Language and Robot Learning 📚🤖 See you in Munich!

Following on from the Open X-Embodiment collaboration, I'm looking forward to speaking at this #CoRL2024 workshop. If we can find efficient ways to transfer policies from your robots to mine, maybe my students will be less tired from having to collect so many demonstrations! 😁

✨New #IROS2024 paper: “Adapting Skills to Novel Grasps: A Self-Supervised Approach” We leverage self-supervised data to adapt skill trajectories to novel object grasp poses. No need for depth, no prior knowledge, no CAD models, no calibration - just RGB images. Here's how👇