Interval Insensitive Loss for Ordinal Classification

Kostiantyn Antoniuk, Vojtech Franc, Vaclav Hlavac
Proceedings of the Sixth Asian Conference on Machine Learning, PMLR 39:189-204, 2015.

Abstract

We address a problem of learning ordinal classifier from partially annotated examples. We introduce an interval-insensitive loss function to measure discrepancy between predictions of an ordinal classifier and a partial annotation provided in the form of intervals of admissible labels. The proposed interval-insensitive loss is an instance of loss functions previously used for learning of different classification models from partially annotated examples. We propose several convex surrogates of the interval-insensitive loss which can be efficiently optimized by existing solvers. Experiments on standard benchmarks and a real-life application show that learning ordinal classifiers from partially annotated examples is competitive to the so-far used methods learning from the complete annotation.

Cite this Paper


BibTeX
@InProceedings{pmlr-v39-antoniuk14, title = {Interval Insensitive Loss for Ordinal Classification}, author = {Antoniuk, Kostiantyn and Franc, Vojtech and Hlavac, Vaclav}, booktitle = {Proceedings of the Sixth Asian Conference on Machine Learning}, pages = {189--204}, year = {2015}, editor = {Phung, Dinh and Li, Hang}, volume = {39}, series = {Proceedings of Machine Learning Research}, address = {Nha Trang City, Vietnam}, month = {26--28 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v39/antoniuk14.pdf}, url = {https://proceedings.mlr.press/v39/antoniuk14.html}, abstract = {We address a problem of learning ordinal classifier from partially annotated examples. We introduce an interval-insensitive loss function to measure discrepancy between predictions of an ordinal classifier and a partial annotation provided in the form of intervals of admissible labels. The proposed interval-insensitive loss is an instance of loss functions previously used for learning of different classification models from partially annotated examples. We propose several convex surrogates of the interval-insensitive loss which can be efficiently optimized by existing solvers. Experiments on standard benchmarks and a real-life application show that learning ordinal classifiers from partially annotated examples is competitive to the so-far used methods learning from the complete annotation.} }
Endnote
%0 Conference Paper %T Interval Insensitive Loss for Ordinal Classification %A Kostiantyn Antoniuk %A Vojtech Franc %A Vaclav Hlavac %B Proceedings of the Sixth Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Dinh Phung %E Hang Li %F pmlr-v39-antoniuk14 %I PMLR %P 189--204 %U https://proceedings.mlr.press/v39/antoniuk14.html %V 39 %X We address a problem of learning ordinal classifier from partially annotated examples. We introduce an interval-insensitive loss function to measure discrepancy between predictions of an ordinal classifier and a partial annotation provided in the form of intervals of admissible labels. The proposed interval-insensitive loss is an instance of loss functions previously used for learning of different classification models from partially annotated examples. We propose several convex surrogates of the interval-insensitive loss which can be efficiently optimized by existing solvers. Experiments on standard benchmarks and a real-life application show that learning ordinal classifiers from partially annotated examples is competitive to the so-far used methods learning from the complete annotation.
RIS
TY - CPAPER TI - Interval Insensitive Loss for Ordinal Classification AU - Kostiantyn Antoniuk AU - Vojtech Franc AU - Vaclav Hlavac BT - Proceedings of the Sixth Asian Conference on Machine Learning DA - 2015/02/16 ED - Dinh Phung ED - Hang Li ID - pmlr-v39-antoniuk14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 39 SP - 189 EP - 204 L1 - http://proceedings.mlr.press/v39/antoniuk14.pdf UR - https://proceedings.mlr.press/v39/antoniuk14.html AB - We address a problem of learning ordinal classifier from partially annotated examples. We introduce an interval-insensitive loss function to measure discrepancy between predictions of an ordinal classifier and a partial annotation provided in the form of intervals of admissible labels. The proposed interval-insensitive loss is an instance of loss functions previously used for learning of different classification models from partially annotated examples. We propose several convex surrogates of the interval-insensitive loss which can be efficiently optimized by existing solvers. Experiments on standard benchmarks and a real-life application show that learning ordinal classifiers from partially annotated examples is competitive to the so-far used methods learning from the complete annotation. ER -
APA
Antoniuk, K., Franc, V. & Hlavac, V.. (2015). Interval Insensitive Loss for Ordinal Classification. Proceedings of the Sixth Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 39:189-204 Available from https://proceedings.mlr.press/v39/antoniuk14.html.

Related Material