An Enhanced Span-based Decomposition Method for Few-Shot Sequence Labeling
This work addresses few-shot sequence labeling for tasks like named entity recognition and slot filling, offering a novel approach that is incremental over prior metric-based methods.
The paper tackles the problem of few-shot sequence labeling by proposing an enhanced span-based decomposition method (ESD) that formulates it as a span-level matching problem, achieving new state-of-the-art results on benchmarks like FewNERD and SNIPS with improved robustness in nested and noisy scenarios.
Few-Shot Sequence Labeling (FSSL) is a canonical paradigm for the tagging models, e.g., named entity recognition and slot filling, to generalize on an emerging, resource-scarce domain. Recently, the metric-based meta-learning framework has been recognized as a promising approach for FSSL. However, most prior works assign a label to each token based on the token-level similarities, which ignores the integrality of named entities or slots. To this end, in this paper, we propose ESD, an Enhanced Span-based Decomposition method for FSSL. ESD formulates FSSL as a span-level matching problem between test query and supporting instances. Specifically, ESD decomposes the span matching problem into a series of span-level procedures, mainly including enhanced span representation, class prototype aggregation and span conflicts resolution. Extensive experiments show that ESD achieves the new state-of-the-art results on two popular FSSL benchmarks, FewNERD and SNIPS, and is proven to be more robust in the nested and noisy tagging scenarios. Our code is available at https://github.com/Wangpeiyi9979/ESD.