Retrieval-augmented Multi-label Text Classification
This addresses the problem of handling infrequent labels in multi-label classification for domains like legal and biomedical texts, though it is incremental as it builds on standard Transformer architectures.
The paper tackled multi-label text classification with large, skewed label sets by using retrieval augmentation to improve sample efficiency, resulting in substantial performance gains on infrequent labels, especially in low-resource and long-document scenarios.
Multi-label text classification (MLC) is a challenging task in settings of large label sets, where label support follows a Zipfian distribution. In this paper, we address this problem through retrieval augmentation, aiming to improve the sample efficiency of classification models. Our approach closely follows the standard MLC architecture of a Transformer-based encoder paired with a set of classification heads. In our case, however, the input document representation is augmented through cross-attention to similar documents retrieved from the training set and represented in a task-specific manner. We evaluate this approach on four datasets from the legal and biomedical domains, all of which feature highly skewed label distributions. Our experiments show that retrieval augmentation substantially improves model performance on the long tail of infrequent labels especially so for lower-resource training scenarios and more challenging long-document data scenarios.