One Day Meeting: Human Robot Interaction and World Models

Wednesday 9 June 2027

Chairs: Filippo Vella (Consiglio Nazionale delle Ricerche)

Queries? contact the Meeting's Organiser Andrew Gilbert here

Invited Speakers

• Prof Angelo Cangelosi, University of Manchester — Professor of Machine Learning & Robotics
• Prof Yannis Demiris, Imperial College London — Professor in Human-Centred Robotics

Call for Presentations

The Human robot interaction hinges on the creation of a synergy between human and an embodied artificial intelligence. When humans and robots share a physical space, their collaboration requires more than coordinated actions: it calls for shared goals, mutual understanding, and planning that account for each partner’s capabilities, intentions and constraints

Successful collaboration requires each partner to understand what the other can do and how their actions may unfold. Robots must therefore consider both their own actions and those of their human partners, anticipating how these actions influence one another and adapting as tasks and environments change. Recent advances in deep learning and cognitive robotics—including world models, vision-language models, and vision-language-action models—offer promising approaches to predicting future states and supporting this adaptive coordination.

World models are increasingly being explored as a means of representing complex environments, anticipating the consequences of actions, and guiding decisions under uncertainty. Whether learned explicitly or encoded implicitly, these models offer opportunities to move beyond static representations toward systems that can respond to evolving situations. Their integration with models that connect language, perception, and action raises important questions about how robots can understand human behaviour and participate effectively in shared activities.

How should these models evolve to support human–robot collaboration? Can structured predictive representations and joint representations of language, vision, and action help robots reason about human intentions, anticipate their partners’ needs, and adapt to changing tasks? How can these capabilities be evaluated in realistic interactions?

We invite contributions that investigate these questions through theoretical, methodological, and empirical work. Topics of interest include:

• World models for the outcomes of joint activities.
• Vision-language and vision-language-action models for collaborative robotics.
• Representations of shared goals, partner capabilities, intentions, and uncertainty.
• Human-aware planning, coordination, and adaptation in dynamic environments.
• Learning and updating models through interaction with human partners.
• Evaluation methods and benchmarks for predictive accuracy, adaptability, and collaboration quality.

Our aim is to bring together researchers exploring how predictive and large models can support more capable, responsive, and context-sensitive human–robot interaction.

Presentations can be either published work, or ongoing research.

The deadline for submitting a Expression of Interest to Present is the 16th April 2027

Meeting Location

The meeting will take place at:

British Computer Society (BCS), 25 Copthall Avenue, London EC2R 7BP

Registration

We keep the cost of attending these events as low as possible to ensure no barriers from the whole computer vision community attending. The registration costs are as follows

• All Attendees: £30 Including lunch and refreshments for the day