Robustness Quantification for Discriminative Models: a New Robustness Metric and its Application to Dynamic Classifier Selection

Rodrigo F L Lassance, Jasper De Bock
Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, PMLR 337:3262-3273, 2026.

Abstract

Among the different possible strategies for evaluating the reliability of individual predictions of classifiers, robustness quantification stands out as a method that evaluates how much uncertainty a classifier could cope with before changing its prediction. However, its applicability is more limited than some of its alternatives, since it requires the use of generative models and restricts the analyses either to specific model architectures or discrete features. In this work, we propose a new robustness metric applicable to any probabilistic discriminative classifier and any type of features. We demonstrate that this new metric is capable of distinguishing between reliable and unreliable predictions, and use this observation to develop new strategies for dynamic classifier selection.

Cite this Paper

BibTeX
@InProceedings{pmlr-v337-lassance26a, title = {Robustness Quantification for Discriminative Models: a New Robustness Metric and its Application to Dynamic Classifier Selection}, author = {Lassance, Rodrigo F L and De Bock, Jasper}, booktitle = {Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence}, pages = {3262--3273}, year = {2026}, editor = {Perković, Emilija and Malinsky, Daniel}, volume = {337}, series = {Proceedings of Machine Learning Research}, month = {17--21 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v337/main/assets/lassance26a/lassance26a.pdf}, url = {https://proceedings.mlr.press/v337/lassance26a.html}, abstract = {Among the different possible strategies for evaluating the reliability of individual predictions of classifiers, robustness quantification stands out as a method that evaluates how much uncertainty a classifier could cope with before changing its prediction. However, its applicability is more limited than some of its alternatives, since it requires the use of generative models and restricts the analyses either to specific model architectures or discrete features. In this work, we propose a new robustness metric applicable to any probabilistic discriminative classifier and any type of features. We demonstrate that this new metric is capable of distinguishing between reliable and unreliable predictions, and use this observation to develop new strategies for dynamic classifier selection.} }
Endnote
%0 Conference Paper %T Robustness Quantification for Discriminative Models: a New Robustness Metric and its Application to Dynamic Classifier Selection %A Rodrigo F L Lassance %A Jasper De Bock %B Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Emilija Perković %E Daniel Malinsky %F pmlr-v337-lassance26a %I PMLR %P 3262--3273 %U https://proceedings.mlr.press/v337/lassance26a.html %V 337 %X Among the different possible strategies for evaluating the reliability of individual predictions of classifiers, robustness quantification stands out as a method that evaluates how much uncertainty a classifier could cope with before changing its prediction. However, its applicability is more limited than some of its alternatives, since it requires the use of generative models and restricts the analyses either to specific model architectures or discrete features. In this work, we propose a new robustness metric applicable to any probabilistic discriminative classifier and any type of features. We demonstrate that this new metric is capable of distinguishing between reliable and unreliable predictions, and use this observation to develop new strategies for dynamic classifier selection.
APA
Lassance, R.F.L. & De Bock, J.. (2026). Robustness Quantification for Discriminative Models: a New Robustness Metric and its Application to Dynamic Classifier Selection. Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 337:3262-3273 Available from https://proceedings.mlr.press/v337/lassance26a.html.

Related Material