Conformal prediction in learning under privileged information paradigm with applications in drug discovery

Niharika Gauraha, Lars Carlsson, Ola Spjuth
; Proceedings of the Seventh Workshop on Conformal and Probabilistic Prediction and Applications, PMLR 91:147-156, 2018.

Abstract

This paper explores conformal prediction in the learning under privileged information (LUPI) paradigm. We use the SVM$+$ realization of LUPI in an inductive conformal predictor, and apply it to the MNIST benchmark dataset and three datasets in drug discovery. The results show that using privileged information produces valid models and improves efficiency compared to standard SVM, however the improvement varies between the tested datasets and is not substantial in the drug discovery applications. More importantly, using SVM$+$ in a conformal prediction framework enables valid prediction intervals at specified significance levels.

Cite this Paper


BibTeX
@InProceedings{pmlr-v91-gauraha18a, title = {Conformal prediction in learning under privileged information paradigm with applications in drug discovery}, author = {Niharika Gauraha and Lars Carlsson and Ola Spjuth}, booktitle = {Proceedings of the Seventh Workshop on Conformal and Probabilistic Prediction and Applications}, pages = {147--156}, year = {2018}, editor = {Alex Gammerman and Vladimir Vovk and Zhiyuan Luo and Evgueni Smirnov and Ralf Peeters}, volume = {91}, series = {Proceedings of Machine Learning Research}, month = {11--13 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v91/gauraha18a/gauraha18a.pdf}, url = {http://proceedings.mlr.press/v91/gauraha18a.html}, abstract = {This paper explores conformal prediction in the learning under privileged information (LUPI) paradigm. We use the SVM$+$ realization of LUPI in an inductive conformal predictor, and apply it to the MNIST benchmark dataset and three datasets in drug discovery. The results show that using privileged information produces valid models and improves efficiency compared to standard SVM, however the improvement varies between the tested datasets and is not substantial in the drug discovery applications. More importantly, using SVM$+$ in a conformal prediction framework enables valid prediction intervals at specified significance levels.} }
Endnote
%0 Conference Paper %T Conformal prediction in learning under privileged information paradigm with applications in drug discovery %A Niharika Gauraha %A Lars Carlsson %A Ola Spjuth %B Proceedings of the Seventh Workshop on Conformal and Probabilistic Prediction and Applications %C Proceedings of Machine Learning Research %D 2018 %E Alex Gammerman %E Vladimir Vovk %E Zhiyuan Luo %E Evgueni Smirnov %E Ralf Peeters %F pmlr-v91-gauraha18a %I PMLR %J Proceedings of Machine Learning Research %P 147--156 %U http://proceedings.mlr.press %V 91 %W PMLR %X This paper explores conformal prediction in the learning under privileged information (LUPI) paradigm. We use the SVM$+$ realization of LUPI in an inductive conformal predictor, and apply it to the MNIST benchmark dataset and three datasets in drug discovery. The results show that using privileged information produces valid models and improves efficiency compared to standard SVM, however the improvement varies between the tested datasets and is not substantial in the drug discovery applications. More importantly, using SVM$+$ in a conformal prediction framework enables valid prediction intervals at specified significance levels.
APA
Gauraha, N., Carlsson, L. & Spjuth, O.. (2018). Conformal prediction in learning under privileged information paradigm with applications in drug discovery. Proceedings of the Seventh Workshop on Conformal and Probabilistic Prediction and Applications, in PMLR 91:147-156

Related Material