Workshop Scope
Machine Learning has revolutionized the world, and has already made a profound impact on our perception of algorithms. More specifically, in recent years, the new area of Algorithms with Predictions has emerged at the intersection of Theoretical Computer Science and ML, which studies the interplay between ML and the design/analysis of algorithms with strict, provable performance guarantees. Its aim is to address fundamental questions related to modeling, performance evaluation and theoretical analysis, from the point of view of both possibility and impossibility results on the theoretical performance. Examples of such directions include:
- How can one leverage (possibly imperfect) predictions, generated by machine learning approaches, in a robust manner, to obtain near-optimal performance when the predictions are accurate, while maintaining worst-case guarantees of classic algorithms?
- How can we model prediction error in ways that lead to a viable and valuable theoretical analysis?
- When predictions are associated with costs, how does an algorithm decide on the ideal times to query a prediction?
- Which tools from ML can be applied in the analysis of algorithms with predictions?
Workshop Organizers:
- Spyros Angelopoulos, CNRS and ILLS Montreal, Canada
- Antonios Antoniadis, University of Twente, The Netherlands
- Marek Eliáš, Bocconi University, Milan, Italy
- Lene Monrad Favrholdt, University of Southern Denmark
- Nicole Megow, University of Bremen, Germany
Invited speakers:
- Sami Davies, Simons Institue for the Theory of Computing and Department of Electrical Engineering and Computer Sciences, UC Berkeley
- Sebastian Forster, Department of Computer Science, University of Salzburg
- Adam Polak, Department of Computing Sciences, Bocconi University
Important Dates:
- submission deadline: May 2nd, 2025
- notification: May 9th, 2025
- Workshop: July 7th, 2025