An Investigation into Automatically Captured Autobiographical Metadata, and the Support for Autobiographical Narrative Generation

FACULTY OF ENGINEERING, SCIENCE AND MATHEMATICS SCHOOL OF ELECTRONICS AND COMPUTER SCIENCE INTELLIGENCE, AGENTS, MULTIMEDIA GROUP A mini-thesis submitted for transfer from MPhil to PhD by Mischa M Tuffield Personal information and the act of publishing multimedia artifacts to the World Wide Web is becoming more and more observable. This report presents an infrastructure for the capturing and exploitation of personal metadata to drive research into context aware systems. I aim to expose ongoing research in the areas of capture of personal experiences, context aware systems, multimedia annotation systems, narrative generation, and that of Semantic Web enabling technologies. This report details the work underway towards the goal of creating a multidomain contextual log, and is followed by a discussion of how this work is being used to drive the development of a multimedia asset management system infrastructure. Practical work already completed as part of this project is then described, focusing on the areas of data integration, infrastructure, and personal photo collection annotation, all in the Semantic Web context. The future plan draws together the work completed and presents how the autobiographical context captured will be evaluated to provide insight into the utility of the metadata harnessed to aid human memory management. This is followed by the proposed evaluation of how suitable semi-automatically captured metadata can be to facilitate the construction of a narrative of a person’s observable life.