Multi-class classification in nonparametric active learning

Boris Ndjia Njike, Xavier Siebert
Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, PMLR 151:7124-7162, 2022.

Abstract

Several works have recently focused on nonparametric active learning, especially in the binary classification setting under Hölder smoothness assumptions on the regression function. These works have highlighted the benefit of active learning by providing better rates of convergence compared to the passive counterpart. In this paper, we extend these results to multiclass classification under a more general smoothness assumption, which takes into account a broader class of underlying distributions. We present a new algorithm called MKAL for multiclass K-nearest neighbors active learning, and prove its theoretical benefits. Additionally, we empirically study MKAL on several datasets and discuss its merits and potential improvements.

Cite this Paper

BibTeX
@InProceedings{pmlr-v151-ndjia-njike22a, title = { Multi-class classification in nonparametric active learning }, author = {Ndjia Njike, Boris and Siebert, Xavier}, booktitle = {Proceedings of The 25th International Conference on Artificial Intelligence and Statistics}, pages = {7124--7162}, year = {2022}, editor = {Camps-Valls, Gustau and Ruiz, Francisco J. R. and Valera, Isabel}, volume = {151}, series = {Proceedings of Machine Learning Research}, month = {28--30 Mar}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v151/ndjia-njike22a/ndjia-njike22a.pdf}, url = {https://proceedings.mlr.press/v151/ndjia-njike22a.html}, abstract = { Several works have recently focused on nonparametric active learning, especially in the binary classification setting under Hölder smoothness assumptions on the regression function. These works have highlighted the benefit of active learning by providing better rates of convergence compared to the passive counterpart. In this paper, we extend these results to multiclass classification under a more general smoothness assumption, which takes into account a broader class of underlying distributions. We present a new algorithm called MKAL for multiclass K-nearest neighbors active learning, and prove its theoretical benefits. Additionally, we empirically study MKAL on several datasets and discuss its merits and potential improvements. } }
Endnote
%0 Conference Paper %T Multi-class classification in nonparametric active learning %A Boris Ndjia Njike %A Xavier Siebert %B Proceedings of The 25th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2022 %E Gustau Camps-Valls %E Francisco J. R. Ruiz %E Isabel Valera %F pmlr-v151-ndjia-njike22a %I PMLR %P 7124--7162 %U https://proceedings.mlr.press/v151/ndjia-njike22a.html %V 151 %X Several works have recently focused on nonparametric active learning, especially in the binary classification setting under Hölder smoothness assumptions on the regression function. These works have highlighted the benefit of active learning by providing better rates of convergence compared to the passive counterpart. In this paper, we extend these results to multiclass classification under a more general smoothness assumption, which takes into account a broader class of underlying distributions. We present a new algorithm called MKAL for multiclass K-nearest neighbors active learning, and prove its theoretical benefits. Additionally, we empirically study MKAL on several datasets and discuss its merits and potential improvements.
APA
Ndjia Njike, B. & Siebert, X.. (2022). Multi-class classification in nonparametric active learning . Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 151:7124-7162 Available from https://proceedings.mlr.press/v151/ndjia-njike22a.html.

Related Material