Don't Blame the Data, Blame the Model: Understanding Noise and Bias When Learning from Subjective Annotations
This addresses the issue of learning from subjective annotations for researchers and practitioners in machine learning, though it appears incremental as it builds on prior work about raw annotations.
The paper tackles the problem of models performing poorly on high-disagreement data in subjective annotation tasks by showing that conventional aggregated-label models have low confidence on such instances, and it demonstrates that using Multiple Ground Truth approaches improves confidence for these instances.
Researchers have raised awareness about the harms of aggregating labels especially in subjective tasks that naturally contain disagreements among human annotators. In this work we show that models that are only provided aggregated labels show low confidence on high-disagreement data instances. While previous studies consider such instances as mislabeled, we argue that the reason the high-disagreement text instances have been hard-to-learn is that the conventional aggregated models underperform in extracting useful signals from subjective tasks. Inspired by recent studies demonstrating the effectiveness of learning from raw annotations, we investigate classifying using Multiple Ground Truth (Multi-GT) approaches. Our experiments show an improvement of confidence for the high-disagreement instances.