Partial-Label Learning with a Reject Option
This addresses the issue of incorrect predictions in partial-label learning for real-world applications where ambiguous labels are common, offering a method to reject unsure predictions to mitigate severe consequences.
The paper tackles the problem of training classifiers with ambiguously labeled data by proposing a risk-consistent nearest-neighbor-based partial-label learning algorithm with a reject option, which shows the best trade-off between the number and accuracy of non-rejected predictions in experiments on artificial and real-world datasets.
In real-world applications, one often encounters ambiguously labeled data, where different annotators assign conflicting class labels. Partial-label learning allows training classifiers in this weakly supervised setting, where state-of-the-art methods already show good predictive performance. However, even the best algorithms give incorrect predictions, which can have severe consequences when they impact actions or decisions. We propose a novel risk-consistent nearest-neighbor-based partial-label learning algorithm with a reject option, that is, the algorithm can reject unsure predictions. Extensive experiments on artificial and real-world datasets show that our method provides the best trade-off between the number and accuracy of non-rejected predictions when compared to our competitors, which use confidence thresholds for rejecting unsure predictions. When evaluated without the reject option, our nearest-neighbor-based approach also achieves competitive prediction performance.