Networked systems are ubiquitous in modern science. Consequently, performing optimization and learning tasks over networks is of utmost importance in several research fields including signal processing, machine learning, optimization, control, biology, economics, computer and social sciences. The combined interplay among distributed processing of information, self-organization, and adaptation is the essential feature of a network designed to perform real-time optimization and learning tasks. A network typically constrains the interaction among individual components: Agents are linked together through a sparse (possibly time-varying) communication topology, and cooperate with each other by relying solely on in-network processing and local sharing of information, in order to accomplish an assigned inferential objective. On the other side, a network may also define structural relationships of data. For instance, high-dimensional data collected over networks (such as social, economic, energy, transportation, telecommunication, biological, to name a few) are naturally modeled as signals defined over graphs, where the graph typically accounts for the topology of the irregular structure of the data. The need for novel analysis tools for graph signals has led to the emergence of the field of graph signal processing, which has been catalyzed by the numerous potential applications such as big data mining, biological data processing, topological data analysis, etc. The goal of this special issue is to gather the latest research efforts toward developing the methodologies necessary to: a) endow networks with distributed optimization, learning, and adaptation capabilities; b) process and analyze signals defined over graphs; b) exploit, disclose, or learn complex relationships and/or patterns hidden in the data collected over networked systems. Therefore, this call for papers encourages submissions related (but not limited to) the following topics of interest.
Edited by: Paolo Di Lorenzo, Antonio G. Marques, Gonzalo Mateos