Deep Affix Features Improve Neural Named Entity Recognizers

Proceedings of the 7th Joint Conference on Lexical and Computational Semantics (*SEM), pages 167–172 New Orleans, June 5-6, 2018. c⃝2018 Association for Computational Linguistics Deep Affix Features Improve Neural Named Entity Recognizers Vikas Yadav†, Rebecca Sharp‡, Steven Bethard† †School of Information, ‡ Dept. of Computer Science University of Arizona, Tucson, AZ, USA {vikasy,bsharp,bethard}@email.arizona.edu Abstract We propose a practical model for named en- tity recognition (NER) that combines word and character-level information with a specific learned representation of the prefixes and suf- fixes of the word. We apply this approach to multilingual and multi-domain NER and show that it achieves state of the art results on the CoNLL 2002 Spanish and Dutch and CoNLL 2003 German NER datasets, consis- tently achieving 1.5-2.3 percent over the state of the art without relying on any dictionary fea- tures. Additionally, we show improvement on SemEval 2013 task 9.1 DrugNER, achieving state of the art results on the MedLine dataset and the second best results overall (-1.3% from state of the art). We also establish a new bench- mark on the I2B2 2010 Clinical NER dataset with 84.70 F-score. 1 Introduction Named entity recognition (NER), or identifying the specific named entities (eg. person, location, organization etc) in a text, is a precursor to other in- formation extraction tasks such as event extraction. The oldest and perhaps most common approach to NER is based on dictionary lookups, and indeed, when the resources are available, this is very useful (e.g., Uzuner et al., 2011). However, hand-crafting these lexicons is time-consuming and expensive and so these resources are often either unavailable or sparse for many domains and languages. Neural network (NN) approaches to NER, on the other hand, do not necessitate these resources, and additionally do not require complex feature engi- neering, which can also be very costly and may not port well from domain to domain and language to language. Commonly, these NN architectures for NER include a learned representation of individual words as well as an encoding of the word’s char- acters. However, neither of these representations makes explicit use of the semantics of sub-word units, i.e., morphemes. Here we propose a simple neural network archi- tecture that learns a custom representation for af- fixes, allowing for a richer semantic representation of words and allowing the model to better approxi- mate the meaning of words not seen during train- ing1. While a full morphological analysis might bring further benefits, to ease re-implementation we take advantage of the Zipfian distribution of lan- guage and focus here on a simple approximation of morphemes as high-frequency prefixes and suffixes. Our approach thus requires no language-specific affix lexicon or morphological tools. Our contributions are: 1. We propose a simple yet robust extension of current neural NER approaches that allows us to learn a representation for prefixes and suffixes of words. We employ an inexpensive and language-independent method to approxi- mate affixes of a given language using n-gram frequencies. This extension is able to be ap- plied directly to new languages …

Content truncated to 500 words