Incorporating Structural Penalties in Multi-label Conformal Prediction

Kostas Katsios, Harris Papadopoulos
Proceedings of the Fourteenth Symposium on Conformal and Probabilistic Prediction with Applications, PMLR 266:230-249, 2025.

Abstract

We propose two structural penalties for the Label-Powerset Split Conformal Prediction framework in multi-label learning. Building on our previously proposed Mahalanobis non- conformity measure, we add penalties that favour label-sets similar to previously observed ones in terms of Hamming distance and cardinality. The resulting nonconformity measure steers prediction regions toward label-sets that are both plausible and compact. Experiments on three public datasets (Emotions, PlantPseAAC, Yeast) show an average of $30%$ reduction in prediction region size for Emotions, $82%$ for PlantPseAAC and $39%$ for Yeast, compared to the Mahalanobis baseline.

Cite this Paper

BibTeX
@InProceedings{pmlr-v266-katsios25a, title = {Incorporating Structural Penalties in Multi-label Conformal Prediction}, author = {Katsios, Kostas and Papadopoulos, Harris}, booktitle = {Proceedings of the Fourteenth Symposium on Conformal and Probabilistic Prediction with Applications}, pages = {230--249}, year = {2025}, editor = {Nguyen, Khuong An and Luo, Zhiyuan and Papadopoulos, Harris and Löfström, Tuwe and Carlsson, Lars and Boström, Henrik}, volume = {266}, series = {Proceedings of Machine Learning Research}, month = {10--12 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v266/main/assets/katsios25a/katsios25a.pdf}, url = {https://proceedings.mlr.press/v266/katsios25a.html}, abstract = {We propose two structural penalties for the Label-Powerset Split Conformal Prediction framework in multi-label learning. Building on our previously proposed Mahalanobis non- conformity measure, we add penalties that favour label-sets similar to previously observed ones in terms of Hamming distance and cardinality. The resulting nonconformity measure steers prediction regions toward label-sets that are both plausible and compact. Experiments on three public datasets (Emotions, PlantPseAAC, Yeast) show an average of $30%$ reduction in prediction region size for Emotions, $82%$ for PlantPseAAC and $39%$ for Yeast, compared to the Mahalanobis baseline.} }
Endnote
%0 Conference Paper %T Incorporating Structural Penalties in Multi-label Conformal Prediction %A Kostas Katsios %A Harris Papadopoulos %B Proceedings of the Fourteenth Symposium on Conformal and Probabilistic Prediction with Applications %C Proceedings of Machine Learning Research %D 2025 %E Khuong An Nguyen %E Zhiyuan Luo %E Harris Papadopoulos %E Tuwe Löfström %E Lars Carlsson %E Henrik Boström %F pmlr-v266-katsios25a %I PMLR %P 230--249 %U https://proceedings.mlr.press/v266/katsios25a.html %V 266 %X We propose two structural penalties for the Label-Powerset Split Conformal Prediction framework in multi-label learning. Building on our previously proposed Mahalanobis non- conformity measure, we add penalties that favour label-sets similar to previously observed ones in terms of Hamming distance and cardinality. The resulting nonconformity measure steers prediction regions toward label-sets that are both plausible and compact. Experiments on three public datasets (Emotions, PlantPseAAC, Yeast) show an average of $30%$ reduction in prediction region size for Emotions, $82%$ for PlantPseAAC and $39%$ for Yeast, compared to the Mahalanobis baseline.
APA
Katsios, K. & Papadopoulos, H.. (2025). Incorporating Structural Penalties in Multi-label Conformal Prediction. Proceedings of the Fourteenth Symposium on Conformal and Probabilistic Prediction with Applications, in Proceedings of Machine Learning Research 266:230-249 Available from https://proceedings.mlr.press/v266/katsios25a.html.

Related Material