Learning from Temporal Data (LearnTeD)
Temporal information is all around us. Numerous important fields, including weather and climate, ecology, transport, urban computing, bioinformatics, medicine, and finance, routinely work with temporal data. Temporal data present a number of new challenges, including increased dimensionality, drifts, complex behavior in terms of long-term interdependence, and temporal sparsity, to mention a few. Hence, learning from temporal data requires specialized strategies that are different from those used for static data. Continuous cross-domain knowledge exchange is required since many of these difficulties cut over the lines separating various fields. This special session aims to integrate the research on learning from temporal data from various areas and to synthesize new concepts based on statistical analysis, time series analysis, graph analysis, signal processing, and machine learning.