วีดีโอ

Authors: Norman Di Palo and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published at: RSS 2024 Paper: arxiv.org/pdf/2403.19578 Webpage: www.robot-learning.uk/keypoint-action-tokens Abstract: We show that off-the-shelf text-based Transformers, with no additional training, can perform few-shot in-context visual imitation learning, mapping visual observations to act...

Authors: Georgios Papagiannis and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published at: CoRL 2024 Paper: arxiv.org/pdf/2410.19693 Webpage: www.robot-learning.uk/miles Abstract: Data collection in imitation learning often requires significant, laborious human supervision, such as numerous demonstrations, and/or frequent environment resets for methods that inco...

Authors: Georgios Papagiannis, Kamil Dreczkowski, Vitalis Vosylius, and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published at: IROS 2024 Paper: arxiv.org/pdf/2408.00178v1 Webpage: www.robot-learning.uk/adapting-skills Abstract: In this paper, we study the problem of adapting manipulation trajectories involving grasped objects (e.g. tools) defined for a single ...

Authors: Teyun Kwon, Norman Di Palo, and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published in: RA-Letters 2024 Paper: arxiv.org/pdf/2310.11604 Webpage: www.robot-learning.uk/language-models-trajectory-generators Abstract: Large Language Models (LLMs) have recently shown promise as high-level planners for robots when given access to a selection of low-level sk...

The Robot Learning Seminar Series is a regular series of in-person seminars at Imperial College London, hosted by Edward Johns, Director of the Robot Learning Lab. For more information, please visit www.robot-learning.uk/seminar-series. Speaker: Markus Wulfmeier (Google DeepMind) Title: Reinforcement Learning in the Age of Large Data Date: Thursday 20th June 2024 Abstract: Recent, vast progress...

The Robot Learning Seminar Series is a regular series of in-person seminars at Imperial College London, hosted by Edward Johns, Director of the Robot Learning Lab. For more information, please visit www.robot-learning.uk/seminar-series. Speaker: Nathan Lepora (University of Bristol) Title: Tactile Robot Dexterity Date: Wednesday 24th April 2024 Abstract: In this talk, I discuss how we can make ...

Authors: Ivan Kapelyukh, Yifei Ren, Ignacio Alzugaray, and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published at: ICRA 2024 Paper: arxiv.org/pdf/2312.04533 Webpage: www.robot-learning.uk/dream2real Abstract: We introduce Dream2Real, a robotics framework which integrates vision-language models (VLMs) trained on 2D data into a 3D object rearrangement pipeline. T...

Authors: Norman Di Palo and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published at: ICRA 2024 Paper: arxiv.org/pdf/2402.13181.pdf Webpage: www.robot-learning.uk/dinobot Abstract: We propose DINOBot, a novel imitation learning framework for robot manipulation, which leverages the image-level and pixel-level capabilities of features extracted from Vision Transfor...

The Robot Learning Seminar Series is a regular series of in-person seminars at Imperial College London, hosted by Edward Johns, Director of the Robot Learning Lab. For more information, please visit www.robot-learning.uk/seminar-series. Speaker: Amanda Prorok (University of Cambridge) Title: Graph Neural Network Based Interaction Models for Collaborative Control in Multi-Robot Systems Date: Wed...

A quick summary of the key papers in 2023 from The Robot Learning Lab at Imperial College London. For further information on these papers, please visit www.robot-learning.uk/research.

Authors: Norman Di Palo and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published in: RA-Letters (accepted December 2023) Paper: arxiv.org/pdf/2312.12345.pdf Webpage: www.robot-learning.uk/retrieval-alignment-replay Abstract: Imitation learning with visual observations is notoriously inefficient when addressed with end-to-end behavioural cloning methods. In this ...

The Robot Learning Seminar Series is a regular series of in-person seminars at Imperial College London, hosted by Edward Johns, Director of the Robot Learning Lab. For more information, please visit www.robot-learning.uk/seminar-series. Speaker: Jamie Shotton (Wayve) Title: Frontiers in Embodied AI for Autonomous Driving Date: Wednesday 15th November 2023 Abstract: Over the last decade, fundame...

Authors: Vitalis Vosylius and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published at: CoRL 2023 Paper: openreview.net/pdf?id=CnKf9TyYtf2 Webpage: www.robot-learning.uk/implicit-graph-alignment Abstract: Consider the following problem: given a few demonstrations of a task across a few different objects, how can a robot learn to perform that same task on new, pre...

Authors: Pietro Vitiello, Kamil Dreczkowski, and Edward Johns Institution: The Robot Learning Lab at Imperial College London Published at: CoRL 2023 Paper: drive.google.com/file/d/1HyB-PrYjibp8Q3Yf2TEPfJQPtKFphAnj/view?usp=sharing Webpage: www.robot-learning.uk/pose-estimation-perspective Abstract: In this paper, we study imitation learning under the challenging setting of: (1) only a single de...

Martin Riedmiller - The Robot Learning Seminar Series

DALL-E-Bot: Introducing Web-Scale Diffusion Models to Robotics

Dimitrios Kanoulas - The Robot Learning Seminar Series

Ingmar Posner - The Robot Learning Seminar Series

Subramanian Ramamoorthy - The Robot Learning Seminar Series

Edward Johns - Annual Review 2022

Where To Start? Transferring Simple Skills to Complex Environments

Demonstrate Once, Imitate Immediately (DOME): Learning Visual Servoing for One-Shot Imitation

Edward Johns - Annual Review 2021

Learning Multi-Stage Tasks with One Demonstration via Self-Replay

Learning Eye-in-Hand Camera Calibration from a Single Image

My House, My Rules: Learning Tidying Preferences with Graph Neural Networks

Hybrid ICP

Coarse-to-Fine for Sim-to-Real: Sub-Millimetre Precision Across Wide Task Spaces

Coarse-to-Fine Imitation Learning: Robot Manipulation from a Single Demonstration