Cross-Lingual Cross-Domain Nested Named Entity Evaluation on English Web Texts

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 1808–1815 August 1–6, 2021. ©2021 Association for Computational Linguistics 1808 Cross-Lingual Cross-Domain Nested Named Entity Evaluation on English Web Texts Barbara Plank IT University of Copenhagen Department of Computer Science Rued Langgaards Vej 7, 2300 København S bplank@gmail.com Abstract Named Entity Recognition (NER) is a key Natural Language Processing task. However, most existing work on NER targets flat named entities (NEs) and ignores the recognition of nested structures, where entities can be en- closed within other NEs. Moreover, evaluation of Nested Named Entity Recognition (NNER) across domains remains challenging, mainly due to the limited availability of datasets. To address these gaps, we present EWT-NNER, a dataset covering five web domains annotated for nested named entities on top of the English Web Treebank (EWT). We present the corpus and an empirical evaluation, including trans- fer results from German and Danish. EWT- NNER is annotated for four major entity types, including suffixes for derivational entity mark- ers and partial named entities, spanning a total of 12 classes. We envision the public release of EWT-NNER to encourage further research on nested NER, particularly on cross-lingual cross-domain evaluation. 1 Introduction Named Entity Recognition (NER) is the task of finding and classifying named entities in text, such as locations, organizations, and person names. It is a key task in Natural Language Processing (NLP), and an important step for downstream applications like relation extraction, co-reference resolution and question answering. The task has received a sub- stantial amount of attention. However, tools and existing benchmarks largely focus on flat, coarse- grained entities and single-domain evaluation. Flat, coarse-grained entities however eschew semantic distinctions which can be important in downstream applications (Ringland et al., 2019). Examples include embedded locations (‘New York Times’), entities formed via derivation (‘Italian cui- sine’) and tokens which are in part named entities (‘the Chicago-based company’). Figure 1: Domain overlap between target (x-axis) and source training (y-axis) domains (DE: German, DA: Danish, EN: proposed dataset EWT-NNER). Research interest on methods to handle nested entities is increasing (Katiyar and Cardie, 2018). However, there is a lack of datasets, particularly resources which cover multiple target domains. To facilitate research on cross-domain nested NER, we introduce a new layer on top of the En- glish Web Treebank (EWT), manually annotated for NNER. The corpus spans five web domains, four major named entity types, enriched with suf- fixes marking derivations and partial NEs. Fig- ure 1 shows the domain overlap in terms of word types. Besides providing in-language benchmark results, EWT-NNER enables research on cross- lingual transfer from German and Danish. Contributions The main contributions are: i) We introduce EWT-NNER, a corpus for nested NER over five web domains. ii) A report on cross-lingual and in-language baselines. Our results highlight the challenges of processing web texts, and the need for research on cross-lingual cross-domain NNER. 2 Related Work Nested NER Much research has been devoted to flat Named Entity Recognition, with a long tradition of shared tasks (Grishman and …

Content truncated to 500 words