Yibo Yan
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 article, 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.
- [1] . 2017. Enriching word vectors with subword information. Trans. Assoc. Computat. Ling. 5 (2017), 135–146.Google Scholar
Cross Ref
- [2] . 2005. A shortest path dependency kernel for relation extraction. In HLT-EMNLP. 724–731.Google Scholar
- [3] . 2018. Adversarial transfer learning for Chinese named entity recognition with self-attention mechanism. In EMNLP. 182–192.Google Scholar
- [4] . 2013. Named entity recognition with bilingual constraints. In NAACL. 52–62.Google Scholar
- [5] . 2006. Chinese named entity recognition with conditional probabilistic models. In SIGHAN. 173–176.Google Scholar
- [6] . 2020. Delinquent events prediction in temporal networked-guarantee loans. IEEE Trans. Neural Netw. Learn. Syst. 34, 4 (2020).Google Scholar
- [7] . 2016. Named entity recognition with bidirectional LSTM-CNNs. Trans. Assoc. Computat. Ling. 4 (2016), 357–370.Google ScholarCross Ref
- [8] . 2019. A neural multi-digraph model for Chinese NER with gazetteers. In ACL. 1462–1467.Google Scholar
- [9] . 2016. Character-based LSTM-CRF with radical-level features for Chinese named entity recognition. In NLPCC. 239–250.Google Scholar
- [10] Thomas Elsken, Jan Hendrik Metzen, and Frank Hutter. 2019. Neural architecture search: A survey. J. Mach. Learn. Res. 20, 1 (2019), 1997–2017.Google Scholar
- [11] . 2013. Paraphrase-driven learning for open question answering. In ACL. 1608–1618.Google Scholar
- [12] . 2020. Fusing global domain information and local semantic information to classify financial documents. In CIKM. 2413–2420.Google Scholar
- [13] Jianping Gou, Baosheng Yu, Stephen J. Maybank, and Dacheng Tao. 2021. Knowledge distillation: A survey. Int. J. Comput. Vis. 129, 6 (2021), 1789–1819.Google Scholar
- [14] . 2019. CNN-based Chinese NER with lexicon rethinking. In IJCAI. 4982–4988.Google Scholar
- [15] . 2019. A lexicon-based graph neural network for Chinese NER. In EMNLP-IJCNLP. 1039–1049.Google Scholar
- [16] . 2015. Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding. arXiv preprint arXiv:1510.00149 (2015).Google Scholar
- [17] . 2017. F-score driven max margin neural network for named entity recognition in Chinese social media. In EACL. 713–718.Google Scholar
- [18] . 2017. A unified model for cross-domain and semi-supervised named entity recognition in Chinese social media. In AAAI. 3216–3222.Google Scholar
- [19] . 2015. Bidirectional LSTM-CRF models for sequence tagging. arXiv preprint arXiv:1508.01991 (2015).Google Scholar
- [20] . 2018. TETRIS: Tile-matching the tremendous irregular sparsity. In NIPS. 4119–4129.Google Scholar
- [21] . 2006. The Third International Chinese Language Processing Bakeoff: Word segmentation and named entity recognition. In SIGHAN. 108–117.Google Scholar
- [22] . 2020. FLAT: Chinese NER using flat-lattice transformer. In ACL. 6836–6842.Google Scholar
- [23] . 2020. F-HMTC: Detecting financial events for investment decisions based on neural hierarchical multi-label text classification. In IJCAI. 4490–4496.Google Scholar
- [24] . 2019. An encoding strategy based word-character LSTM for Chinese NER. In NAACL. 2379–2389.Google Scholar
- [25] . 2016. Multi-prototype chinese character embedding. In LREC. 855–859.Google Scholar
- [26] . 2019. Porous lattice-based transformer encoder for Chinese NER. arXiv preprint arXiv:1911.02733 (2019).Google Scholar
- [27] Andriy Mnih and Russ R. Salakhutdinov. 2007. Probabilistic matrix factorization. In NIPS. 1257–1264.Google Scholar
- [28] . 2020. Simplify the usage of lexicon in Chinese NER. In ACL. 5951–5960.Google Scholar
- [29] . 2015. Named entity recognition for Chinese social media with jointly trained embeddings. In EMNLP. 548–554.Google Scholar
- [30] . 2016. Improving named entity recognition for Chinese social media with word segmentation representation learning. In ACL. 149–155.Google Scholar
- [31] . 2021. ZH-NER: Chinese named entity recognition with adversarial multi-task learning and self-attentions. In DASFAA. 1–8.Google Scholar
- [32] . 2019. Leverage lexical knowledge for Chinese named entity recognition via collaborative graph network. In EMNLP-IJCNLP. 3821–3831.Google Scholar
- [33] . 2019. Distantly supervised entity relation extraction with adapted manual annotations. In AAAI. 7039–7046.Google Scholar
- [34] . 2017. Soft weight-sharing for neural network compression. arXiv preprint arXiv:1702.04008 (2017).Google Scholar
- [35] . 2013. Effective bilingual constraints for semi-supervised learning of named entity recognizers. In AAAI. 52–62.Google Scholar
- [36] . 2011. OntoNotes 4.0. Linguistic Data Consortium LDC2011T03 (2011).Google Scholar
- [37] . 2019. Neural Chinese named entity recognition via CNN-LSTM-CRF and joint training with word segmentation. In WWW. 3342–3348.Google Scholar
- [38] . 2018. Improving neural fine-grained entity typing with knowledge attention. In AAAI. 5997–6004.Google Scholar
- [39] . 2019. TENER: Adapting transformer encoder for name entity recognition. arXiv preprint arXiv:1911.04474 (2019).Google Scholar
- [40] . 2018. Five-stroke based CNN-BiRNN-CRF network for Chinese named entity recognition. In NLPCC. 184–195.Google Scholar
- [41] . 2016. Combining discrete and neural features for sequence labeling. In CICLing. 140–154.Google Scholar
- [42] . 2018. Hybrid semi-Markov CRF for neural sequence labeling. In ACL. 235–240.Google Scholar
- [43] . 2006. Word segmentation and named entity recognition for SIGHAN Bakeoff3. In SIGHAN. 158–161.Google Scholar
- [44] . 2018. Chinese NER using lattice LSTM. In ACL. 1554–1564.Google Scholar
- [45] . 2013. Chinese named entity recognition via joint identification and categorization. Chinese J. Electron. 22, 2 (2013), 225–230.Google Scholar
- [46] . 2022. Improving Chinese named entity recognition by large-scale syntactic dependency graph. IEEE/ACM Trans. Audio, Speech Lang. Process. 30 (2022), 979–991.Google ScholarDigital Library
- [47] . 2019. CAN-NER: Convolutional attention network for Chinese named entity recognition. In NAACL. 3384–3393.Google Scholar
Index Terms
- Adversarial Multi-task Learning for Efficient Chinese Named Entity Recognition
Recommendations
ZH-NER: Chinese Named Entity Recognition with Adversarial Multi-task Learning and Self-Attentions
Database Systems for Advanced ApplicationsAbstractNER is challenging because of the semantic ambiguities in academic literature, especially for non-Latin languages. Besides, recognizing Chinese named entities needs to consider word boundary information, as words contained in Chinese texts are not ...
Learning multilingual named entity recognition from Wikipedia
We automatically create enormous, free and multilingual silver-standard training annotations for named entity recognition (ner) by exploiting the text and structure of Wikipedia. Most ner systems rely on statistical models of annotated data to identify ...
Semi-joint labeling for chinese named entity recognition
AIRS'08: Proceedings of the 4th Asia information retrieval conference on Information retrieval technologyNamed entity recognition (NER) is an essential component of text mining applications. In Chinese sentences, words do not have delimiters; thus, incorporating word segmentation information into an NER model can improve its performance. Based on the ...
Comments