Uncertain data in learning: challenges and opportunities

Sébastien Destercke
Proceedings of the Eleventh Symposium on Conformal and Probabilistic Prediction with Applications, PMLR 179:322-332, 2022.

Abstract

Dealing with uncertain data in statistical estimation problems or in machine learning is not really a new issue. However, such uncertainty has so far mostly been modelled either as sets, being called for instance coarse data or partial labels, or as probability distributions over data values, being called for instance soft labels. Integrating this uncertainty in the learning process can be challenging, but also rewarding, as it can improve both the quality of the made predictions as well as our understanding of the obtained model. Within this setting, rich uncertainty models generalizing both probabilities and sets offer both new challenges and opportunities, and I will summarise some of them in this short note.

Cite this Paper

BibTeX
@InProceedings{pmlr-v179-destercke22a, title = {Uncertain data in learning: challenges and opportunities}, author = {Destercke, S\'{e}bastien}, booktitle = {Proceedings of the Eleventh Symposium on Conformal and Probabilistic Prediction with Applications}, pages = {322--332}, year = {2022}, editor = {Johansson, Ulf and Boström, Henrik and An Nguyen, Khuong and Luo, Zhiyuan and Carlsson, Lars}, volume = {179}, series = {Proceedings of Machine Learning Research}, month = {24--26 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v179/destercke22a/destercke22a.pdf}, url = {https://proceedings.mlr.press/v179/destercke22a.html}, abstract = {Dealing with uncertain data in statistical estimation problems or in machine learning is not really a new issue. However, such uncertainty has so far mostly been modelled either as sets, being called for instance coarse data or partial labels, or as probability distributions over data values, being called for instance soft labels. Integrating this uncertainty in the learning process can be challenging, but also rewarding, as it can improve both the quality of the made predictions as well as our understanding of the obtained model. Within this setting, rich uncertainty models generalizing both probabilities and sets offer both new challenges and opportunities, and I will summarise some of them in this short note. } }
Endnote
%0 Conference Paper %T Uncertain data in learning: challenges and opportunities %A Sébastien Destercke %B Proceedings of the Eleventh Symposium on Conformal and Probabilistic Prediction with Applications %C Proceedings of Machine Learning Research %D 2022 %E Ulf Johansson %E Henrik Boström %E Khuong An Nguyen %E Zhiyuan Luo %E Lars Carlsson %F pmlr-v179-destercke22a %I PMLR %P 322--332 %U https://proceedings.mlr.press/v179/destercke22a.html %V 179 %X Dealing with uncertain data in statistical estimation problems or in machine learning is not really a new issue. However, such uncertainty has so far mostly been modelled either as sets, being called for instance coarse data or partial labels, or as probability distributions over data values, being called for instance soft labels. Integrating this uncertainty in the learning process can be challenging, but also rewarding, as it can improve both the quality of the made predictions as well as our understanding of the obtained model. Within this setting, rich uncertainty models generalizing both probabilities and sets offer both new challenges and opportunities, and I will summarise some of them in this short note.
APA
Destercke, S.. (2022). Uncertain data in learning: challenges and opportunities. Proceedings of the Eleventh Symposium on Conformal and Probabilistic Prediction with Applications, in Proceedings of Machine Learning Research 179:322-332 Available from https://proceedings.mlr.press/v179/destercke22a.html.

Related Material