Active Learning for Name Entity Recognition with External Knowledge

Named Entity Recognition (NER) is an important task in knowledge extraction, which targets extracting structural information from unstructured text. To fully employ the prior-knowledge of pre-trained language models, some research works formulate NER into machine reading comprehension form (MRC-form) to enhance their model generalization capability commonsense knowledge. However, this transformation still faces data-hungry issue with limited training data for specific tasks. address low-resource NER, we introduce a method named active multi-task-based (AMT-NER), two-stage multi-task learning model. Specifically, A module first introduced AMT-NER improve its representation Then, strategy proposed optimize learning. An associated Natural Language Inference (NLI) also employed further. More importantly, introduces module, uncertainty selective, actively filter help learn efficiently. Besides, find different external supportive under pipelines improves performance differently Extensive experiments are performed show superiority our method, proves findings that introduction significant and effective MRC-form