Reducing Bias in Pre-Trained Models by Tuning While Penalizing Change

Niklas Penzel, Gideon Stein, Joachim Denzler

2024

Abstract

Deep models trained on large amounts of data often incorporate implicit biases present during training time. If later such a bias is discovered during inference or deployment, it is often necessary to acquire new data and retrain the model. This behavior is especially problematic in critical areas such as autonomous driving or medical decision-making. In these scenarios, new data is often expensive and hard to come by. In this work, we present a method based on change penalization that takes a pre-trained model and adapts the weights to mitigate a previously detected bias. We achieve this by tuning a zero-initialized copy of a frozen pre-trained network. Our method needs very few, in extreme cases only a single, examples that contradict the bias to increase performance. Additionally, we propose an early stopping criterion to modify baselines and reduce overfitting. We evaluate our approach on a well-known bias in skin lesion classification and three other datasets from the domain shift literature. We find that our approach works especially well with very few images. Simple fine-tuning combined with our early stopping also leads to performance benefits for a larger number of tuning samples.

Paper Citation

in Harvard Style

Penzel N., Stein G. and Denzler J. (2024). Reducing Bias in Pre-Trained Models by Tuning While Penalizing Change. In Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP; ISBN 978-989-758-679-8, SciTePress, pages 90-101. DOI: 10.5220/0012345800003660

in Bibtex Style

@conference{visapp24,
author={Niklas Penzel and Gideon Stein and Joachim Denzler},
title={Reducing Bias in Pre-Trained Models by Tuning While Penalizing Change},
booktitle={Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP},
year={2024},
pages={90-101},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012345800003660},
isbn={978-989-758-679-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP
TI - Reducing Bias in Pre-Trained Models by Tuning While Penalizing Change
SN - 978-989-758-679-8
AU - Penzel N.
AU - Stein G.
AU - Denzler J.
PY - 2024
SP - 90
EP - 101
DO - 10.5220/0012345800003660
PB - SciTePress