SeedBERT: Recovering Annotator Rating Distributions from an Aggregated Label
This addresses the issue of annotator disagreement in subjective tasks like affective computing, which is incremental as it builds on pre-trained models to handle existing data limitations.
The authors tackled the problem of subjective machine learning tasks where datasets have only aggregated labels, by proposing SeedBERT to recover annotator rating distributions from a single label. Their method showed substantial performance gains on downstream subjective tasks compared to existing models.
Many machine learning tasks -- particularly those in affective computing -- are inherently subjective. When asked to classify facial expressions or to rate an individual's attractiveness, humans may disagree with one another, and no single answer may be objectively correct. However, machine learning datasets commonly have just one "ground truth" label for each sample, so models trained on these labels may not perform well on tasks that are subjective in nature. Though allowing models to learn from the individual annotators' ratings may help, most datasets do not provide annotator-specific labels for each sample. To address this issue, we propose SeedBERT, a method for recovering annotator rating distributions from a single label by inducing pre-trained models to attend to different portions of the input. Our human evaluations indicate that SeedBERT's attention mechanism is consistent with human sources of annotator disagreement. Moreover, in our empirical evaluations using large language models, SeedBERT demonstrates substantial gains in performance on downstream subjective tasks compared both to standard deep learning models and to other current models that account explicitly for annotator disagreement.