Signal Processing for graphs

An increasing number of data streams can be interpreted as structured according to a graph of similarities or relationships between nodes.
Such data is either collected over the nodes or over the edges of a graph, which may be only partially known. The latter type of data is relational and includes, for instance, transactions between people in a social network. The former type of data is attributional and includes, for example, volume per second of traffic from different IP addresses. This talk will provide an overview of some of the major trends in graph-structured modeling of data with an emphasis on signal processing applications.