Uncovering Non-verbal Semantic Aspects Of Collaborative Meetings: Iterative Design And Evaluation Of The Meeting Miner

摘要

This paper describes the process of iterative design and comprehensive evaluation of the Meeting Miner, a tool for browsing of recorded multimedia meetings. The Meeting Miner provides access to the speech content of recordings by tracking non-verbal interaction events collected in real time during online collaborative meeting activities. It emphasises semantic relationships between speech and discrete actions performed during the meeting and aggregates information through patterns of co-location and co-occurrence of actions. We report on the experience gained through developing functionality to enhance the user's browsing experience, requirements regarding information feedback and the importance of flexibility in the browsing tool. In particular, we demonstrate how iterative development and evaluation can reveal areas where interface adaptation can play a useful role in enhancing the system's usability.