Learning to benchmark (A. Hero, University of Michigan, USA)

Jul 14
14-07-2021 16:00 Uhr bis 17:00 Uhr
Online

Learning to benchmark

Speaker: Prof. Dr. Alfred Hero
Affiliation: University of Michigan, USA
Zoom link: Meeting-ID: 69989364767, Passcode: 233558

Title: Learning to benchmark
Abstract: We address the problem of learning an achievable lower bound on classification error from a labeled sample. We establish an optimization framework for this meta-learning problem, which we call benchmark learning. Benchmark learning leads to an accurate data-driven predictor of performance of a Bayes optimal classifier without having to construct the classifier and without assuming any parametric model for the data. The resultant predictor can be used to establish whether it is possible to improve classification performance of a specific classifier. It also yields a stopping rule for sequentially trained classifiers. In addition, The talk will cover relevant background, theory, algorithms, and applications of benchmark learning.

https://en.www.math.fau.de/applied-analysis/

Friedrich-Alexander-Universität Erlangen-Nürnberg