Cross-Paradigm Graph Algorithms (Invited Talk)

A goal of the theory of graph algorithms is algorithmic techniques that enable computing devices to process graph data with little resources (time, space, communication overhead, etc.). This led to extensive studies of graph algorithms in various models of computation (sequential algorithms, distributed algorithms, streaming algorithms, etc.) by many sub-communities. Cross-paradigm graph algorithms is an effort to attack the same problem in many models of computation simultaneously, with the goal to generate new insights that may not emerge from the isolated viewpoint of a single model and to ultimately develop techniques that can be used to solve graph problems near-optimally across many models of computation. In this talk, I will discuss some recent advances in graph algorithmic techniques for basic graph problems (e.g. minimum cut, shortest path, and maximum flow) in connection to this research program, especially some insights that led to cross-paradigm algorithms and to answering notorious open questions. No background will be assumed from the audience beyond familiarity with textbook graph algorithms.