Dual IAS Seminar - Associate Professor Sumiko Miyata & Professor Takamichi Miyata

Externally Funded Fellows Associate Professor Sumiko Miyata & Professor Takamichi Miyata each deliver a seminar on their research -

Associate Professor Sumiko Miyata - Incentive-Driven AI Networks for Future Road Safety 

To achieve fully autonomous driving, "cooperative perception" via V2X (Vehicle-to-Everything) is essential for eliminating blind spots and improving recognition accuracy. However, a major barrier to sustainable implementation lies in ensuring "fair incentives" for participants to share data and computational resources. This seminar introduces an AI-driven network framework designed to balance infrastructure efficiency with participant satisfaction. The presentation first covers a reward distribution mechanism based on the game theory concept of "Nucleolus" to minimize user dissatisfaction within the monitoring system and ensure long-term cooperation. Building on this foundation, the discussion addresses essential network mechanisms for "City as a Service," such as high-speed AI processing that optimizes task offloading between edge servers to minimize communication latency. By integrating incentive design with advanced communication control, it is possible to build a reliable social infrastructure that reduces accidents and optimizes urban mobility.

Professor Takamichi Miyata - Multimodal AI that Understands Driver Behaviour without Training Data 

Distracted driving remains a critical safety concern, as even brief lapses in attention can lead to serious traffic collisions. Current supervised learning methods require large, labelled datasets and struggle to generalize, while vision-language model (VLM) based methods enable training-free recognition but tend to capture driver identity rather than actual behaviour. This seminar presents a novel framework that overcomes both limitations. The key innovation lies in decoupling identity-related information from behaviour-related cues, combined with refined textual representations to enhance zero-shot recognition robustness across diverse drivers and environments. By integrating decoupled multimodal representations with a lightweight model architecture, the proposed system achieves practical, scalable performance without relying on extensive labelled data. This approach offers a promising pathway toward reliable driver monitoring systems for real-world deployment.

Arrivals from 11:45 am for a 12:00 noon start. For those joining in-person, lunch will be served after the seminar from 1:00pm.

This event is hybrid format, please use the required booking button at the bottom of the page to choose either in-person or online attendance.
(Please note that in-person spaces are limited and booking is required, so we can manage numbers for catering and also the space in the seminar room)

By booking a place at this event, attendees agree to behave in a respectful manner such that everyone feels comfortable contributing as they wish. The IAS reserves the right to eject anyone who does not abide by this policy.

IAS seminars are typically recorded, minus any Q&A sessions at the end, again to encourage contributions. The recordings are then uploaded to our website on a Fellows bio page and/or Programme page, along with our IAS YouTube Channel. If you are not able to attend a seminar live, please do still register for the online webinar, as we will email everyone who registered to let them know once the recordings are made available.