Ir directamente a la navegación principal Ir directamente a la búsqueda Ir directamente al contenido principal

Crowd learning with candidate labeling: An EM-based solution

  • Basque Center for Applied Mathematics
  • University of the Basque Country

Producción científica: Capítulo del libro/Acta de congresoActa de congresorevisión exhaustiva

1 Cita (Scopus)

Resumen

Crowdsourcing is widely used nowadays in machine learning for data labeling. Although in the traditional case annotators are asked to provide a single label for each instance, novel approaches allow annotators, in case of doubt, to choose a subset of labels as a way to extract more information from them. In both the traditional and these novel approaches, the reliability of the labelers can be modeled based on the collections of labels that they provide. In this paper, we propose an Expectation-Maximization-based method for crowdsourced data with candidate sets. Iteratively the likelihood of the parameters that model the reliability of the labelers is maximized, while the ground truth is estimated. The experimental results suggest that the proposed method performs better than the baseline aggregation schemes in terms of estimated accuracy.

Idioma originalInglés
Título de la publicación alojadaAdvances in Artificial Intelligence - 18th Conference of the Spanish Association for Artificial Intelligence, CAEPIA 2018, Proceedings
EditoresAntonio Gonzalez, Alicia Troncoso, Francisco Herrera, Sergio Damas, Rosana Montes, Sergio Alonso, Oscar Cordon
EditorialSpringer Verlag
Páginas13-23
Número de páginas11
ISBN (versión impresa)9783030003739
DOI
EstadoPublicada - 2018
Publicado de forma externa
Evento18th Conference of the Spanish Association for Artificial Intelligence, CAEPIA 2018 - Granada, Espana
Duración: 23 oct 201826 oct 2018

Serie de la publicación

NombreLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volumen11160 LNAI
ISSN (versión impresa)0302-9743
ISSN (versión digital)1611-3349

Conferencia

Conferencia18th Conference of the Spanish Association for Artificial Intelligence, CAEPIA 2018
País/TerritorioEspana
CiudadGranada
Período23/10/1826/10/18

Huella

Profundice en los temas de investigación de 'Crowd learning with candidate labeling: An EM-based solution'. En conjunto forman una huella única.

Citar esto