Multimodal Phased Transformer for Sentiment Analysis

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 198–207 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 198 TEBNER: Domain Specific Named Entity Recognition with Type Expanded Boundary-aware Network Zheng Fang1,2, Yanan Cao1,2∗, Tai Li3, Ruipeng Jia1,2 Fang Fang1,2, Yanmin Shang1,2 and Yuhai Lu1,2∗ 1Institute of Information Engineering, Chinese Academy of Sciences 2School of Cyber Security, University of Chinese Academy of Sciences 3Bytedance 1,2{fangzheng, caoyanan, jiaruipeng, fangfang0703, shangyanmin luyuhai }@iie.ac.cn, 3litai.vincent@bytedance.com Abstract To alleviate label scarcity in Named Entity Recognition (NER) task, distantly supervised NER methods are widely applied to automati- cally label data and identify entities. Although the human effort is reduced, the generated incomplete and noisy annotations pose new challenges for learning effective neural mod- els. In this paper, we propose a novel dictio- nary extension method which extracts new en- tities through the type expanded model. More- over, we design a multi-granularity boundary- aware network which detects entity bound- aries from both local and global perspectives. We conduct experiments on different types of datasets, the results show that our model out- performs previous state-of-the-art distantly su- pervised systems and even surpasses the super- vised models. 1 Introduction Named Entity Recognition (NER) is the task of de- tecting mentions from text and classifying them into predefined types. It is a fundamental task in the field of natural language processing (NLP), which can facilitate many other tasks, such as en- tity linking (Fang et al., 2020), machine translation (Gekhman et al., 2020), and question answering (Li et al., 2020a). However, most of existing NER methods require large amounts of manually anno- tated texts for training supervised models, which is difficult to implement in the specific domain be- cause domain-expert annotation is expensive and time-consuming. To alleviate the label scarcity problem, distant su- pervision methods (Fries et al., 2017; Shang et al., 2018b; Liang et al., 2020) have been applied to automatically generate labeled data and recognize entities. Given a raw corpus and a dictionary, above methods firstly label entities by exact string match- ing, and then use the annotated dataset to train the ∗Corresponding authors: Yanan Cao and Yuhai Lu. well-designed neural models to recognize entities. Although the human effort is reduced, the labels generated by the string matching method pose two challenges. The first challenge is incomplete annotations. Because most of existing dictionaries have lim- ited coverage on domain entities, just using the given dictionary will make many out-of-dictionary entities unmatched and generate a large number of false-negative labels. By analyzing several commonly used datasets (e.g., BC5CDR, NCBI, MeSH), we find that the original dictionary only covers about 50% of domain entities, which may weaken the performance of subsequent NER model. To increase the number of label entities, previous works (Fries et al., 2017; Liu et al., 2020) attempt to expand the dictionary by heuristic rules. However, these rule-based methods are difficult to migrate to other domains. So, how to extend the dictionary with more general pattern is the first problem we need …

Content truncated to 500 words