Thompson, A; Jagan, K; Sundar, A; Khatry, R; Donlevy, J; Thomas, S; Harris, P (2021) Uncertainty evaluation for machine learning. NPL Report. MS 34

Preview

Text (NPL Report MS 34) MS34.pdf - Published Version Download (2MB) | Preview

Abstract

We explore the challenge of uncertainty evaluation in machine learning regression problems. We begin by outlining how regression problems typically arise in a metrology setting. We then perform a numerical investigation of three methods for performing uncertainty-aware machine learning regression, namely Gaussian Processes (GPs), Monte-Carlo Dropout and Deep Ensembles. Based upon our investigations, we identify a list of metrology requirements that methods for uncertainty evaluation in machine learning regression must satisfy, and we evaluate the three methods against these requirements. We conclude by describing how the requirements give rise naturally to several open challenges which constitute an agenda for future work.

Item Type:	Report/Guide (NPL Report)
NPL Report No.:	MS 34
Subjects:	Mathematics and Scientific Computing > Numerical Computation
Divisions:	Data Science
Identification number/DOI:	10.47120/npl.MS34
Last Modified:	11 Jan 2022 11:38
URI:	https://eprintspublications.npl.co.uk/id/eprint/9306