Successful exploitation of Big Data hinges on innovative Computer Science algorithms for dealing with the so-called 3-V characteristics of Big Data: Velocity, Volume, and Variety. The Velocity characteristic in particular requires so-called dynamic algorithms that can support data analytics in the presence of real-time updates.

While there is disparate work on dynamic analytics in individual areas of Computer Science, this project proposes a uniform framework to dynamically solve diverse analytic problems that recur frequently in Computer Science in areas as diverse as constraint satisfaction, databases, machine learning, matrix operations, probabilistic graphical models inference, and logic.

By studying these dynamic analytics problems in a common framework, and implementing them in open-source software, this project will establish novel dynamic, worst- case-optimal algorithms that are of immediate relevance to the host of problem settings mentioned above.


  • Dan Olteanu, Department of Computer Science, University of Oxford
  • Stijn Vansummeren, Brussels School of Engineering