Learning from biased crowdsourced labeling with deep clustering

Ming Wu,Qianmu Li,Fei Yang,Jing Zhang,Victor S. Sheng,Jun Hou

DOI: https://doi.org/10.1016/j.eswa.2022.118608

IF: 8.5

2023-01-01

Expert Systems with Applications

Abstract:•The phenomenon of biased labeling usually existing in the scenario of crowdsourcing.•Biased labeling is a critical factor that effects label aggregation performance.•Deep clustering estimates the underlying label distribution and detect the bias.