Machine Learning with Known Input Data Uncertainty Measure

Abstract : Uncertainty of the input data is a common issue in machine learning. In this paper we show how one can incorporate knowledge on uncertainty measure regarding particular points in the training set. This may boost up models accuracy as well as reduce overfitting. We show an approach based on the classical training with jitter for Artificial Neural Networks (ANNs). We prove that our method, which can be applied to a wide class of models, is approximately equivalent to generalised Tikhonov regularisation learning. We also compare our results with some alternative methods. In the end we discuss further prospects and applications.
Type de document :
Communication dans un congrès
Khalid Saeed; Rituparna Chaki; Agostino Cortesi; Sławomir Wierzchoń. 12th International Conference on Information Systems and Industrial Management (CISIM), Sep 2013, Krakow, Poland. Springer, Lecture Notes in Computer Science, LNCS-8104, pp.379-388, 2013, Computer Information Systems and Industrial Management. 〈10.1007/978-3-642-40925-7_35〉
Liste complète des métadonnées

Littérature citée [12 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01496084
Contributeur : Hal Ifip <>
Soumis le : lundi 27 mars 2017 - 11:01:46
Dernière modification le : mardi 28 mars 2017 - 01:07:10
Document(s) archivé(s) le : mercredi 28 juin 2017 - 13:16:31

Fichier

978-3-642-40925-7_35_Chapter.p...
Fichiers produits par l'(les) auteur(s)

Licence


Distributed under a Creative Commons Paternité 4.0 International License

Identifiants

Citation

Wojciech Czarnecki, Igor Podolak. Machine Learning with Known Input Data Uncertainty Measure. Khalid Saeed; Rituparna Chaki; Agostino Cortesi; Sławomir Wierzchoń. 12th International Conference on Information Systems and Industrial Management (CISIM), Sep 2013, Krakow, Poland. Springer, Lecture Notes in Computer Science, LNCS-8104, pp.379-388, 2013, Computer Information Systems and Industrial Management. 〈10.1007/978-3-642-40925-7_35〉. 〈hal-01496084〉

Partager

Métriques

Consultations de la notice

154

Téléchargements de fichiers

599