Learning and Evaluation in Presence of Non-i.i.d. Label Noise

Nico Görnitz, Anne Porbadnigk, Alexander Binder, Claudia Sannelli, Mikio Braun, Klaus-Robert Mueller, Marius Kloft
Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, PMLR 33:293-302, 2014.

Abstract

In many real-world applications, the simplified assumption of independent and identically distributed noise breaks down, and labels can have structured, systematic noise. For example, in brain-computer interface applications, training data is often the result of lengthy experimental sessions, where the attention levels of participants can change over the course of the experiment. In such application cases, structured label noise will cause problems because most machine learning methods assume independent and identically distributed label noise. In this paper, we present a novel methodology for learning and evaluation in presence of systematic label noise. The core of which is a novel extension of support vector data description / one-class SVM that can incorporate latent variables. Controlled simulations on synthetic data and a real-world EEG experiment with 20 subjects from the domain of brain-computer-interfacing show that our method achieves accuracies that go beyond the state of the art.

Cite this Paper

BibTeX
@InProceedings{pmlr-v33-gornitz14, title = {{Learning and Evaluation in Presence of Non-i.i.d. Label Noise}}, author = {Görnitz, Nico and Porbadnigk, Anne and Binder, Alexander and Sannelli, Claudia and Braun, Mikio and Mueller, Klaus-Robert and Kloft, Marius}, booktitle = {Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics}, pages = {293--302}, year = {2014}, editor = {Kaski, Samuel and Corander, Jukka}, volume = {33}, series = {Proceedings of Machine Learning Research}, address = {Reykjavik, Iceland}, month = {22--25 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v33/gornitz14.pdf}, url = {https://proceedings.mlr.press/v33/gornitz14.html}, abstract = {In many real-world applications, the simplified assumption of independent and identically distributed noise breaks down, and labels can have structured, systematic noise. For example, in brain-computer interface applications, training data is often the result of lengthy experimental sessions, where the attention levels of participants can change over the course of the experiment. In such application cases, structured label noise will cause problems because most machine learning methods assume independent and identically distributed label noise. In this paper, we present a novel methodology for learning and evaluation in presence of systematic label noise. The core of which is a novel extension of support vector data description / one-class SVM that can incorporate latent variables. Controlled simulations on synthetic data and a real-world EEG experiment with 20 subjects from the domain of brain-computer-interfacing show that our method achieves accuracies that go beyond the state of the art.} }
Endnote
%0 Conference Paper %T Learning and Evaluation in Presence of Non-i.i.d. Label Noise %A Nico Görnitz %A Anne Porbadnigk %A Alexander Binder %A Claudia Sannelli %A Mikio Braun %A Klaus-Robert Mueller %A Marius Kloft %B Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2014 %E Samuel Kaski %E Jukka Corander %F pmlr-v33-gornitz14 %I PMLR %P 293--302 %U https://proceedings.mlr.press/v33/gornitz14.html %V 33 %X In many real-world applications, the simplified assumption of independent and identically distributed noise breaks down, and labels can have structured, systematic noise. For example, in brain-computer interface applications, training data is often the result of lengthy experimental sessions, where the attention levels of participants can change over the course of the experiment. In such application cases, structured label noise will cause problems because most machine learning methods assume independent and identically distributed label noise. In this paper, we present a novel methodology for learning and evaluation in presence of systematic label noise. The core of which is a novel extension of support vector data description / one-class SVM that can incorporate latent variables. Controlled simulations on synthetic data and a real-world EEG experiment with 20 subjects from the domain of brain-computer-interfacing show that our method achieves accuracies that go beyond the state of the art.
RIS
TY - CPAPER TI - Learning and Evaluation in Presence of Non-i.i.d. Label Noise AU - Nico Görnitz AU - Anne Porbadnigk AU - Alexander Binder AU - Claudia Sannelli AU - Mikio Braun AU - Klaus-Robert Mueller AU - Marius Kloft BT - Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics DA - 2014/04/02 ED - Samuel Kaski ED - Jukka Corander ID - pmlr-v33-gornitz14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 33 SP - 293 EP - 302 L1 - http://proceedings.mlr.press/v33/gornitz14.pdf UR - https://proceedings.mlr.press/v33/gornitz14.html AB - In many real-world applications, the simplified assumption of independent and identically distributed noise breaks down, and labels can have structured, systematic noise. For example, in brain-computer interface applications, training data is often the result of lengthy experimental sessions, where the attention levels of participants can change over the course of the experiment. In such application cases, structured label noise will cause problems because most machine learning methods assume independent and identically distributed label noise. In this paper, we present a novel methodology for learning and evaluation in presence of systematic label noise. The core of which is a novel extension of support vector data description / one-class SVM that can incorporate latent variables. Controlled simulations on synthetic data and a real-world EEG experiment with 20 subjects from the domain of brain-computer-interfacing show that our method achieves accuracies that go beyond the state of the art. ER -
APA
Görnitz, N., Porbadnigk, A., Binder, A., Sannelli, C., Braun, M., Mueller, K. & Kloft, M.. (2014). Learning and Evaluation in Presence of Non-i.i.d. Label Noise. Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 33:293-302 Available from https://proceedings.mlr.press/v33/gornitz14.html.

Related Material