Abstract:Named entity recognition (NER) is a fundamental task for information extraction applications. NER is challenging because of semantic ambiguities in academic literature, especially for non-Latin languages. Besides word semantic information, recognizing Chinese named entities needs to consider word boundary information, as words contained in Chinese texts are not separated with spaces. Leveraging word boundary information could help to determine entity boundaries and thus improve entity recognition performance. In this paper, we propose to combine word boundary information and semantic information for named entity recognition based on multi-task adversarial learning. Specifically, we learn commonly shared boundary information of entities from multiple kinds of tasks, including Chinese word segmentation (CWS), part-of-speech (POS) tagging and entity recognition, with adversarial learning. We learn task-specific semantic information of words from these tasks and combine the learned boundary information with the semantic information to improve entity recognition, with multi-task learning. We then propose a compression method based on improved clustering to accelerate the proposed model. We conduct extensive experiments on four public benchmark datasets and two private datasets, compared with state-of-the-art baseline models, and the experimental results demonstrate that our model achieves considerable performance improvements on various evaluation datasets.

Complex Named Entity Recognition via Deep Multi-task Learning from Scratch.

A Multi-task Framework for Named Entity Recognition

Towards Hierarchical Categorical Named Entity Recognition Via Injective Multi-task Learning Strategy

Named Entity Recognition via Machine Reading Comprehension: A Multi-Task Learning Approach

ZH-NER: Chinese Named Entity Recognition with Adversarial Multi-task Learning and Self-Attentions.

A Multi-task Approach for Machine Reading Comprehension Form Named Entity Recognition Tasks

Adversarial Multi-Task Learning for Efficient Chinese Named Entity Recognition

Cross-type Biomedical Named Entity Recognition with Deep Multi-Task Learning

Multitask Learning for Chinese Named Entity Recognition.

Bidirectional GRU with Multi-Head Attention for Chinese NER

Adversarial Multitask Learning for Technology Entity Recognition

Pronounce Differently, Mean Differently: A Multi-Tagging-scheme Learning Method for Chinese NER Integrated with Lexicon and Phonetic Features

Recognizing Nested Named Entity in Biomedical Texts: A Neural Network Model with Multi-Task Learning

Learning with Joint Cross-Document Information Via Multi-Task Learning for Named Entity Recognition

Enhancing NER with Sentence-Level Entity Detection As an Simple Auxiliary Task

Chinese Named Entity Recognition with a Multi-Phase Model

Chinese named entity recognition: The state of the art

Innovative Deep Neural Network Modeling For Fine-Grained Chinese Entity Recognition

Joint Cross-document Information for Named Entity Recognition with Multi-task Learning

A Double Adversarial Network Model for Multi-Domain and Multi-Task Chinese Named Entity Recognition

Multi-task learning for Chinese clinical named entity recognition with external knowledge