Statistical and Probabilistic Methods in Algorithmic Data Analysis (Dagstuhl Seminar 24391)

Abstract

This report documents the program and the outcomes of Dagstuhl Seminar "Statistical and Probabilistic Methods in Algorithmic Data Analysis" (24391). Modern algorithms for data analysis require the use of advanced probabilistic methods to achieve the necessary scalability and accuracy guarantees. At the same time, modern tasks of knowledge discovery from data require the use of advanced statistics to handle challenges such as the test of multiple hypotheses or dependency structure of the data points, such as in time series or graphs. Probabilistic methods are also at the core of areas of theoretical computer science such as sub-linear algorithms and average-case analysis. The application of these methods requires careful balancing of theoretical and practical considerations, to obtain efficient algorithms for data analysis. The Dagstuhl Seminar focused on statistical and probabilistic methods to develop and analyze useful, scalable algorithms for knowledge discovery from large, rich datasets. Participants from different countries, at different stages of their careers, and from both industry and academia gave talks on the topics of the seminar, usually presenting their own research, either recently published or soon-to-be. There was ample time for socializing, networking, and starting or continuing collaborations, and new results are expected to be published thanks to these collaborations.