Biography
Prof. Ingmar Posner leads the Applied Artificial Intelligence Lab (A2I) at the University of Oxford. A founding director of the Oxford Robotics Institute (ORI), Ingmar’s career is shaped by a vision to create machines that constantly improve through experience, with a core research focus on representation learning and structured world models for robotics and beyond. He co-led the team behind the UK’s first autonomous car, the Oxford RobotCar (now part of the permanent collection of the London Science Museum), and is a co-founder of Oxa, an award-winning provider of autonomous vehicle software. In 2023, the ORI was awarded a Queen’s Anniversary Prize, the highest national honour for UK educational institutions. An ELLIS Fellow, Ingmar currently serves as an Amazon Scholar.
Research Interests
Prof. Posner's research focus spans the algorithmic foundations and real-world applications of AI and reflects his long-standing interest in how intelligent agents can robustly perceive, ground, and interact with complex environments. His work has produced seminal contributions to robot learning, spanning large-scale learning from demonstration, unsupervised learning of scene dynamics, object-centric generative modelling, and 3D object detection. It has been recognised with multiple Best Paper Awards at leading international venues such as ICAPS, IROS, and ISER.
Central to his group's research is the development of structured representation learning and world modelling. Rather than relying on purely task-specific learning, this paradigm is designed to enable machines to capture re-usable, interpretable representations of the world's underlying structure and dynamics. By moving beyond traditional perception boundaries, these models help autonomous systems, and increasingly scientists, reason about complex environments, improving generalisation and interpretability across diverse domains ranging from robot manipulation to the natural sciences.
Recent Publications
An End-to-End Framework for Modeling Pneumatic Soft Robots Based on Differentiable Finite Element Methods
Zhong S, Yao Y, Maiolino P & Posner I (2025), IEEE Robotics and Automation Letters, 10(12), 13003-13010
BibTeX
@article{anendtoendframe-2025/12,
title={An End-to-End Framework for Modeling Pneumatic Soft Robots Based on Differentiable Finite Element Methods},
author={Zhong S, Yao Y, Maiolino P & Posner I},
journal={IEEE Robotics and Automation Letters},
volume={10},
pages={13003-13010},
publisher={Institute of Electrical and Electronics Engineers (IEEE)},
year = "2025"
}
The complexity dynamics of grokking
DeMoss B, Sapora S, Foerster J, Hawes N & Posner I (2025), Physica D Nonlinear Phenomena, 134859
Offline Adaptation of Quadruped Locomotion Using Diffusion Models
O'Mahoney R, Mitchell AL, Yu W, Posner I & Havoutis I (2025), 00, 9974-9980
BibTeX
@inproceedings{offlineadaptati-2025/5,
title={Offline Adaptation of Quadruped Locomotion Using Diffusion Models},
author={O'Mahoney R, Mitchell AL, Yu W, Posner I & Havoutis I},
booktitle={2025 IEEE International Conference on Robotics and Automation (ICRA)},
pages={9974-9980},
year = "2025"
}
Group decision-making in robot teleoperation: two heads are better than one
Nguyen A, Bhattacharyya R, Colombatto C, Fleming S, Posner H et al. (2025), 489-500
BibTeX
@inproceedings{groupdecisionma-2025/3,
title={Group decision-making in robot teleoperation: two heads are better than one},
author={Nguyen A, Bhattacharyya R, Colombatto C, Fleming S, Posner H et al.},
booktitle={IEEE/ACM International Conference on Human-Robot Interaction (HRI 2025)},
pages={489-500},
year = "2025"
}
Group Decision-Making in Robot Teleoperation: Two Heads are Better Than One
Nguyen D-A, Bhattacharyya R, Colombatto C, Fleming S, Posner I et al. (2025), 00, 489-500
BibTeX
@inproceedings{groupdecisionma-2025/3,
title={Group Decision-Making in Robot Teleoperation: Two Heads are Better Than One},
author={Nguyen D-A, Bhattacharyya R, Colombatto C, Fleming S, Posner I et al.},
booktitle={2025 20th ACM/IEEE International Conference on Human-Robot Interaction (HRI)},
pages={489-500},
year = "2025"
}
COMBO-Grasp: learning constraint-based manipulation for bimanual occluded grasping
Yamada J, Mitchell A, Collins J & Posner I (2025)
Joint Decision-Making in Robot Teleoperation: When are Two Heads Better Than One?
Nguyen D-A, Bhattacharyya R, Colombatto C, Fleming S, Posner I et al. (2025)
BibTeX
@misc{jointdecisionma-2025/1,
title={Joint Decision-Making in Robot Teleoperation: When are Two Heads Better
Than One?},
author={Nguyen D-A, Bhattacharyya R, Colombatto C, Fleming S, Posner I et al.},
year = "2025"
}
Vision-Language-Action Models for Robotics: A Review Towards Real-World Applications
Kawaharazuka K, Oh J, Yamada J, Posner I & Zhu Y (2025), IEEE Access, 13, 162467-162504
BibTeX
@article{visionlanguagea-2025/1,
title={Vision-Language-Action Models for Robotics: A Review Towards Real-World Applications},
author={Kawaharazuka K, Oh J, Yamada J, Posner I & Zhu Y},
journal={IEEE Access},
volume={13},
pages={162467-162504},
publisher={Institute of Electrical and Electronics Engineers (IEEE)},
year = "2025"
}
TactGen: tactile sensory data generation via zero-shot sim-to-real transfer
Zhong S, Albini A, Maiolino P & Posner I (2024), IEEE Transactions on Robotics, 41, 1316-1328
BibTeX
@article{tactgentactiles-2024/12,
title={TactGen: tactile sensory data generation via zero-shot sim-to-real transfer},
author={Zhong S, Albini A, Maiolino P & Posner I},
journal={IEEE Transactions on Robotics},
volume={41},
pages={1316-1328},
publisher={Institute of Electrical and Electronics Engineers},
year = "2024"
}
The Complexity Dynamics of Grokking
DeMoss B, Sapora S, Foerster J, Hawes N & Posner I (2024)
BibTeX
@misc{thecomplexitydy-2024/12,
title={The Complexity Dynamics of Grokking},
author={DeMoss B, Sapora S, Foerster J, Hawes N & Posner I},
year = "2024"
}