Presentation on YouTube.

When studying an object, for instance a graph, measuring distances is helpful. In this talk, we will first present a variety of distances that may be used in graphs depending on the context. Indeed, the same measure will unlikely be meaningful for graphs associated diverse situations. For example, the shortest path is well suited for telecommunication networks but may not be used for conceptual graphs (like semantic networks) or in the case of epidemic spreading. From that premise, the need to implement and study distances in AutoGraphiX (AGX) appears. In the second part of the talk, we will present the way distances may be studied in AGX as well as the challenges it involves both from the optimization and data analysis points of view.