Leveraging Context for Adaptive Multimedia Retrieval: A Matter of Control

摘要

Adaptive systems have not been adopted because people prefer stability and predictability for high-level interactions and because such systems have not been robust enough to handle the variety of goals and tasks that people bring to them. Adaptive features at fine grains of activity have been successfully adopted and this paper considers the possibilities of using a variety of contextual features for adaptive multimedia retrieval. It argues that people must gain trust in such systems and be provided with control over the adaptive features within a more general human-computer information retrieval framework.