TY - GEN
T1 - Learning from crowds in multi-dimensional classification domains
AU - Hernández-González, Jerónimo
AU - Inza, Iñaki
AU - Lozano, José A.
PY - 2013
Y1 - 2013
N2 - Learning from crowds is a recently fashioned supervised classification framework where the true/real labels of the training instances are not available. However, each instance is provided with a set of noisy class labels, each indicating the class-membership of the instance according to the subjective opinion of an annotator. The additional challenges involved in the extension of this framework to the multi-label domain are explored in this paper. A solution to this problem combining a Structural EM strategy and the multi-dimensional Bayesian network models as classifiers is presented. Using real multi-label datasets adapted to the crowd framework, the designed experiments try to shed some lights on the limits of learning to classify from the multiple and imprecise information of supervision.
AB - Learning from crowds is a recently fashioned supervised classification framework where the true/real labels of the training instances are not available. However, each instance is provided with a set of noisy class labels, each indicating the class-membership of the instance according to the subjective opinion of an annotator. The additional challenges involved in the extension of this framework to the multi-label domain are explored in this paper. A solution to this problem combining a Structural EM strategy and the multi-dimensional Bayesian network models as classifiers is presented. Using real multi-label datasets adapted to the crowd framework, the designed experiments try to shed some lights on the limits of learning to classify from the multiple and imprecise information of supervision.
KW - Bayesian network models
KW - Learning from crowds
KW - Multi-label classification
KW - Structural EM method
KW - multi-dimensional classification
UR - https://www.scopus.com/pages/publications/84885056288
U2 - 10.1007/978-3-642-40643-0_36
DO - 10.1007/978-3-642-40643-0_36
M3 - Conference proceeding
AN - SCOPUS:84885056288
SN - 9783642406423
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 352
EP - 362
BT - Advances in Artificial Intelligence - 15th Conference of the Spanish Association for Artificial Intelligence, CAEPIA 2013, Proceedings
T2 - 15th Conference of the Spanish Association for Artificial Intelligence, CAEPIA 2013
Y2 - 17 September 2013 through 20 September 2013
ER -