Date: 29 March 2023
Time: 3:00 PM ET (New York Time)
Speaker(s): Dr. Gerald Matz
SPS-DSI Webinar Series: Data SciEnce on GrAphS (DEGAS)
Initiated by the Data Science Initiative of IEEE Signal Processing Society, the DEGAS Webinar Series serves to provide the SP community with updates and advances in learning and inference on graphs. Signal processing and machine learning often deal with data living in regular domains such as space and time.
This webinar series will cover the extension of these methods to network data, including topics such as graph filtering, graph sampling, spectral analysis of network data, graph topology identification, geometric deep learning, and so on. Applications can, for instance, be found in image processing, social networks, epidemics, wireless communications, brain science, recommender systems, and sensor networks.
Webinars are hosted on Zoom, with recordings available on the IEEE SPS YouTube channel following the live event. Each speaker presentation is followed by time for Q&A and discussions.
Gerald Matz received the Dipl.-Ing. (1994) and Dr. techn. (2000) degrees in Electrical Engineering and the Habilitation degree (2004) for Communication Systems from TU Wien, Austria. He currently holds a tenured Associate Professor position with the Institute of Telecommunications at TU Wien. In 2004-2005 he was on leave as an Erwin Schrödinger Fellow with the Laboratoire des Signaux et Systèmes (Supelec, France). In 2007 he was a Visiting Researcher with the Communication Theory Lab at ETH Zurich (Switzerland), and in 2011 he was a Guest Professor with ENSEEIHT (Toulouse, France).
Traditionally, graphs in signal processing and machine learning have been constructed from data using a notion of similarity. Often it is useful to augment such constructions by incorporating the notion of an opposite (different polarity), which results in a (single) dichotomy. Some datasets, however, require multiple dichotomies (i.e., similarity/opposite in multiple data dimensions). In this talk, I will show how to deal with such data sets by using signed graphs and signature graphs. I will argue that for these types of graphs balance theory provides a natural and elegant framework for clustering and classification and I will discuss how to learn the topology of signed and signature graphs from given data.