Personalized Access to Meeting Recordings for Knowledge Transfer and Learning Purposes in Companies
Christian Gütl, Christian Safran · 2006
Meetings increasingly take place in our daily business life. Knowledge workers and mangers spend between 20 und 80 % of their working time in meetings. Consequently a huge amount of knowledge is addressed or even generated in such meetings. In light of that an important but also challenging task is to manage that knowledge as part of the corporate knowledge repository and make it accessible for further usage. The aim of the MISTRAL research project is to deal with those problems. The MISTRAL core system automatically annotates multi-modal meeting recordings on a semantic level, such as information about the attendees of a meeting, topics addressed and speech-to-text transcripts. MISTRAL’s semantic application deals with indexing and retrieval of this semantic information for knowledge transfer and learning purposes. In this paper we want to focus in particular on personalization issues in that context. Our approach enables personalized access to background information, or more concrete meeting information, based on the concept level for a given context, such as a particular learning or working tasks. Based on our Context-sensitive Concept Modeling System enhanced with personalization functionality, an architecture, first implementation results and lessons learned are presented.