Due to the growing presence of ubiquitous data sources and the requirement for shorter and shorter response times, data streaming is steadily becoming the key technology to modern information processing, and this trend is bringing up new challenges. As data analysis moves from simple organisation and classification to content understanding, for instance, the meaning of segmentation becomes more and more complex to define. When does a concept start within a text, and where does it stop? Which objects are interesting in a video, and where are they located? When does a machine start to show a degradation process that will lead to a need for maintenance? Where does a spoken text start to show a certain mood?

This special session aims to explore current topics in data stream understanding, with emphasis on segmentation, the process of identifying meaningful entities within a stream of data. This is often an ill-defined goal, where object boundaries are uncertain and object definition is unclear, subjective, or not given in advance. Notwithstanding this, the task emerges everywhere and can be instantiated in a vast number of vertical application areas.

Organizers