From media crossing to media mining

This paper reviews how the concept of Media Crossing has contributed to the advancement of the application domain of information access and explores directions for a future research agenda. These will include themes that could help to broaden the scope and to incorporate the concept of medium-crossing in a more general approach that not only uses combinations of medium-specific processing, but that also exploits more abstract medium-independent representations, partly based on the foundational work on statistical language models for information retrieval. Three examples of successful applications of media crossing will be presented, with a focus on the aspects that could be considered a first step towards a generalized form of media mining.