Deep neural learning on weighted datasets utilizing label disagreement from crowdsourcing

Experts and crowds can work together to generate high-quality datasets, but such collaboration is limited a large-scale pool of data. In other words, training on dataset depends more crowdsourced datasets with aggregated labels than expert intensively checked labels. However, the amount be used as an objective test build connection between disagreement this paper, we claim that behind label indicates semantics (e.g. ambiguity or difficulty) instance just spam error assessment. We attempt take advantage informativeness assist learning neural networks by computing series measurements incorporating distinct mechanisms. Experiments two demonstrate consideration disagreement, treating instances differently, promisingly result in improved performance.