DaNE: A Named Entity Resource for Danish

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 4597–4604 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 4597 DaNE: A Named Entity Resource for Danish Rasmus Hvingelby1∗, Amalie Brogaard Pauli1∗, Maria Barrett2, Christina Rosted1, Lasse Malm Lidegaard1, Anders Søgaard2 1Alexandra Institute, Denmark 2University of Copenhagen, Denmark Abstract We present a named entity annotation for the Danish Universal Dependencies treebank using the CoNLL-2003 annotation scheme: DaNE. It is the largest publicly available, Danish named entity gold annotation. We evaluate the quality of our annotations intrinsically by double annotating the entire treebank and extrinsically by comparing our annotations to a recently released named entity annotation of the validation and test sections of the Danish Universal Dependencies treebank. We benchmark the new resource by training and evaluating competitive architectures for supervised named entity recognition (NER), including FLAIR, monolingual (Danish) BERT and multilingual BERT. We explore cross-lingual transfer in multilingual BERT from five related languages in zero-shot and direct transfer setups, and we show that even with our modestly-sized training set, we improve Danish NER over a recent cross-lingual approach, as well as over zero-shot transfer from five related languages. Using multilingual BERT, we achieve higher performance by fine-tuning on both DaNE and a larger Bokm˚al (Norwegian) training set compared to only using DaNE. However, the highest performance is achieved by using a Danish BERT fine-tuned on DaNE. Our dataset enables improvements and applicability for Danish NER beyond cross-lingual methods. We employ a thorough error analysis of the predictions of the best models for seen and unseen entities, as well as their robustness on un-capitalized text. The annotated dataset and all the trained models are made publicly available. Keywords: named entity recognition, resource, Danish, cross-lingual transfer 1. Introduction Named entity recognition (NER) is a crucial, yet challeng- ing, component of a wide range of natural language pro- cessing applications, ranging from knowledge base popula- tion and other natural language understanding tasks to pri- vacy protection systems. For Danish, there has not yet been a larger named entity (NE) annotated training set and until very recently; also no test set. Despite the lack of freely available training data, a few NER tools exist for Danish (Bick, 2004; Johannessen et al., 2005; Derczynski, 2019; Al-Rfou et al., 2015) but until Plank (2019), they have not been consistently benchmarked. This study found that Dan- ish NER can benefit from transfer from English. In this paper, we present a new annotated resource for Dan- ish NE, DaNE, which contains annotations of the Danish Universal Dependencies treebank (Johannsen et al., 2015). We describe the annotation process and evaluate the annota- tion quality intrinsically by looking at annotator agreement, and extrinsically by comparing to the recent NE annotation of the validation and test split of the same resource (Plank, 2019). We benchmark our new resource with supervised state-of-the-art NE taggers, both off-the-shelf systems and models that are trained or fine-tuned on our train set. Our train set is smaller than …

Content truncated to 500 words