Simple and Effective Few-Shot Named Entity Recognition with Structured Nearest Neighbor Learning

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 6365–6375, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 6365 Simple and Effective Few-Shot Named Entity Recognition with Structured Nearest Neighbor Learning Yi Yang ASAPP Inc. New York, NY 10007 yyang@asapp.com Arzoo Katiyar ∗ Pennsylvania State University University Park, PA 16802 arzoo@psu.edu Abstract We present a simple few-shot named entity recognition (NER) system based on nearest neighbor learning and structured inference. Our system uses a supervised NER model trained on the source domain, as a feature ex- tractor. Across several test domains, we show that a nearest neighbor classifier in this feature- space is far more effective than the standard meta-learning approaches. We further propose a cheap but effective method to capture the label dependencies between entity tags with- out expensive CRF training. We show that our method of combining structured decoding with nearest neighbor learning achieves state- of-the-art performance on standard few-shot NER evaluation tasks, improving F1 scores by 6% to 16% absolute points over prior meta- learning based systems. 1 Introduction Named entity recognition (NER) aims at identify- ing and categorizing spans of text into a closed set of classes, such as people, organizations, and locations. As a core language understanding task, NER is widely adopted in several domains, such as news (Tjong Kim Sang and De Meulder, 2003), medical (Stubbs and Uzuner, 2015), and social me- dia (Derczynski et al., 2017). However, one of the primary challenges in adapting NER to new domains is the mismatch between the different domain-specific entity types. For example, only two out of the twenty-three entity types annotated in the I2B2 2014 (Stubbs and Uzuner, 2015) data can be found in the OntoNotes 5 (Weischedel et al., 2013) annotations. Unfortunately, obtaining NER annotations for a novel domain can be quite expen- sive, often requiring domain knowledge. Few-shot classification (Vinyals et al., 2016; Bao et al., 2020) models aim at recognizing new classes based on only few labeled examples (support set) ∗Work done at ASAPP Inc. Figure 1: An few-shot NER example. Professions (e.g., ‘minister’ and ‘president’) and dates (e.g., ‘today’ and ‘tomorrow’) are part of the O class. Nearest neigh- bor classifier is better at predicting the O class using an instance-based metric compared to methods using class-based metric. from each class. In the context of NER, these few- shot classification methods can enable rapid build- ing of NER systems for a new domain by labeling only a few examples per entity class. Several previ- ous studies (Fritzler et al., 2019; Hou et al., 2020) propose using prototypical networks (Snell et al., 2017), a popular few-shot classification algorithm, to address the few-shot NER problem. However, these approaches only achieve 10 ∼ 30% F1 scores on average, when transferring knowledge between different NER datasets with one or five shot ex- amples, warranting more effective methods for the problem. The direct adaption of existing few-shot classifi- cation methods to few-shot NER is challenging for two reasons. First, NER is essentially a …

Content truncated to 500 words