Making Learners (More) Monotone

Conference paper (2020)

Authors

T.J. Viering Pattern Recognition and Bioinformatics - EEMCS
A. Mey Interactive Intelligence - EEMCS
M. Loog University of Copenhagen, Pattern Recognition and Bioinformatics - EEMCS
Research Group
Pattern Recognition and Bioinformatics (EEMCS) (TU Delft)
Copyright: © 2020 T.J. Viering, A. Mey, M. Loog
DOI: https://doi.org/10.1007/978-3-030-44584-3_42
More Info
expand_more

Published Date

2020

Language

English

Faculty

Electrical Engineering, Mathematics and Computer Science

Department

Intelligent Systems

Research Group

Pattern Recognition and Bioinformatics

Abstract

Learning performance can show non-monotonic behavior. That is, more data does not necessarily lead to better models, even on average. We propose three algorithms that take a supervised learning model and make it perform more monotone. We prove consistency and monotonicity with high probability, and evaluate the algorithms on scenarios where non-monotone behaviour occurs. Our proposed algorithm MTHT makes less than 1% non-monotone decisions on MNIST while staying competitive in terms of error rate compared to several baselines. Our code is available at https://github.com/tomviering/monotone.