Automated, scalable systems would reveal and help exploit the deeper meanings in scientific data, especially in biomedical engineering, telecommunications, geospatial exploration, and climate and Earth ecosystem modeling.

ecent progress in scientific and engi-neering applications has accumulatedhuge volumes of high-dimensionaldata, stream data, unstructured andsemi-structured data, and spatial andtemporal data. Highly scalable andsophisticated data mining tools forsuch applications represent one of themost active research frontiers in datamining. Here, we outline the related challenges inseveral emerging domains.