Events

Date: 23 October 2025   Time: 14:00 - 15:00

Location: Hybrid: Seminar Room MB-503, School of Mathematical Sciences, QMUL, or via the Teams link below

In this talk, we present findings from an empirical study of popular dimension estimators. The first half of the talk will give an overview of common mathematical approaches in dimensions estimation (including topological data analysis based approaches), while the second half will focus on analysing the performance of estimators under different noise levels and sampling densities. We discover that many estimators require careful choices of hyperparameters for accurate estimation, and often fail on datasets with moderately high dimensions. Drawing from our empirical results, we give some practical advice for using dimension estimators and interpreting their outputs.

This talk is based on joint work with James Binnie, Paweł Dłotko, John Harvey and Jakub Malinowski in arXiv:2507.13887.