The Common RNN Based Models with Sequence Tagging

摘要

Sequence tagging is a very important and basic research direction in the field of natural language processing. The performance of many top-level NLP tasks largely depends on the accuracy and reliability of sequence tagging. Correlation model has made a significant progress in this field. At present, the focus of research is how to make better use of the correlation and timing between data to sequence tagging tasks on the premise of ensuring robustness, so as to achieve further performance. In this paper, we summarize the highlights and breakthrough research results and research models in the field of sequence tagging in recent years, mainly from the models, data sets, parameters and performance. At the same time, we summarize the existing problems and look forward to the future.