Paper Information | SBC BOOKS

Fill in your paper information

Language (*)

Title (*)

Keywords (*)

Abstract (*)

Arquiteturas neurais baseadas em transformers tornaram-se o principal componente de vários métodos do estado-da-arte em tarefas de processamento de linguagem natural, tais como Reconhecimento de Entidades Nomeadas e Extração de Relações (REN+ER). Como essas arquiteturas baseiam-se em aspectos semânticos de sequências de palavras, elas podem não funcionar na identificação e delimitação de entidades nomeadas quando há pouco contexto semântico associado, tais como entidades compostas por dígitos e pontuações apenas (e.g., números de CPF) e entidades com nomes compostos. Neste artigo, são propostas novas técnicas de reforço contextual e delimitação de entidades baseadas em pré- e pós-processamento de dados para enriquecer o contexto semântico, melhorando assim um método do estado-da-arte para REN+RE, o SpERT (Span-Based Entity and Relation Transformer). Tais técnicas foram avaliadas usando dados reais de diários oficiais e de processos judiciais. Os resultados mostram que, quando aplicadas em conjunto, as estratégias de pré- e pós-processamento levam a ganhos significativos na efetividade de REN+ER.

Pages (*)

File Link

English Information

Title

Keywords

Abstract

Transformer architectures have become the main component of various state-of-the-art methods for natural language processing tasks, such as Named Entity Recognition and Relation Extraction (NER+RE). As these architectures rely on semantic aspects of word sequences, they may fail to accurately identify and delimit entity spans when there is little semantic context surrounding the named entities. This is the case of entities composed by digits and punctuation only, such as IDs and phone numbers, as well as long composed names. In this paper, we propose new techniques for contextual reinforcement and entity delimitation based on pre- and post-processing techniques to provide a richer semantic context, improving SpERT, a state-of-the-art Span-based Entity and Relation Transformer. We evaluate our strategies using real data from public administration documents and court lawsuits. Our results show that our pre- and post-processing strategies, when used co-jointly, allows significant improvements on NER+ER effectiveness.

(*) To change the order drag the item to the new position.

Authors

#	Name
1	Fabiano Belém(fmuniz@dcc.ufmg.br)
2	Marcelo Ganem(masganem@dcc.ufmg.br)
3	Celso França(celsofranca@dcc.ufmg.br)
4	Marcos Carvalho(marcoscarvalho@dcc.ufmg.br)
5	Alberto Laender(laender@dcc.ufmg.br)
6	Marcos Gonçalves(mgoncalv@dcc.ufmg.br)

(*) To change the order drag the item to the new position.

Reference

#	Reference
1	Brunner, U. & Stockinger, K. (2020). Entity Matching with Transformer Architectures - A Step Forward in Data Integration. In International Conference on Extending Database Technology, pages 463–473
2	Caputo, A., Basile, P., & Semeraro, G. (2009). Boosting a Semantic Search Engine by Named Entities. In Foundations of Intelligent Systems, pages 241–250
3	Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Conference of the of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4171–4186
4	Eberts, M. & Ulges, A. (2020). Span-based Joint Entity and Relation Extraction with Transformer Pretraining. In 24th European Conference on Artificial Intelligence, pages 2006–2013
5	Eberts, M. & Ulges, A. (2021). An End-to-end Model for Entity-level Relation Extraction using Multiinstance Learning. In Association for Computational Linguistics, pages 3650–3660
6	Finkel, J. R., Grenager, T., & Manning, C. (2005). Incorporating non-local information into information extraction systems by Gibbs sampling. In Annual Meeting of the Association for Computational Linguistics, pages 363–370
7	Fu, J., Huang, X., & Liu, P. (2021). SpanNER: Named Entity Re-/Recognition as Span Prediction. In Annual Meeting of the Association for Computational Linguistics, pages 7183–7195
8	Liu, C., Fan, H., & Liu, J. (2021). Span-based nested named entity recognition with pretrained language model. In Jensen, C. S., Lim, E.-P., Yang, D.-N., Lee, W.-C., Tseng, V. S., Kalogeraki, V., Huang, J.-W., & Shen, C.-Y., editors, Database Systems for Advanced Applications, pages 620–628
9	Luz de Araujo, P. H., de Campos, T. E., de Oliveira, R. R. R., Stauffer, M., Couto, S., & Bermejo, P. (2018). LeNER-Br: a dataset for named entity recognition in Brazilian legal text. In International Conference on the Computational Processing of Portuguese (PROPOR), pages 313–323
10	Niu, F., Zhang, C., R´e, C., & Shavlik, J. W. (2012). DeepDive: Web-scale Knowledge-base Construction using Statistical Learning and Inference. VLDS, 12:25–28
11	Patil, N., Patil, A., & Pawar, B. (2020). Named Entity Recognition using Conditional Random Fields. Procedia Computer Science, 167:1181–1188
12	Silva, L., Canalle, G. K., Salgado, A. C., L´oscio, B., & Moro, M. (2019). Uma An´alise Experimental do Impacto da Selec¸ ˜ao de Atributos em Processos de Resoluc¸ ˜ao de Entidades. In SBBD, pages 37–48
13	Wang, T., Zhao, X., Lv, Q., Hu, B., & Sun, D. (2021). Density weighted diversity based query strategy for active learning. In IEEE International Conference on Computer Supported Cooperative Work in Design (CSCWD), pages 156–161.
14	Zhang, S., He, L., Vucetic, S., & Dragut, E. (2018). Regular Expression Guided Entity Mention Mining from Noisy Web Data. In Empirical Methods in Natural Language Processing, pages 1991–2000

Paper Registration

Fill in your paper information

English Information

Authors

Reference