Language model based on deep learning network for biomedical named entity recognition.

Area of Science:

• Biomedical text mining
• Natural Language Processing
• Bioinformatics

Background:

• Biomedical Named Entity Recognition (BioNER) is crucial for extracting information from biomedical literature.
• Deep learning methods show promise but struggle with polysemous entities and limited training data.

Purpose of the Study:

• To develop a novel multi-task learning framework for BioNER to overcome challenges of word ambiguity and data scarcity.
• To improve the accuracy and robustness of biomedical entity recognition.

Main Methods:

• Proposed a multi-task learning framework based on BiLSTM-CRF architecture, integrating a language model for differential context encoding.
• Utilized dynamic word vectors to disambiguate polysemous biological entities.
• Employed multi-task learning to share information across different entity types, enhancing recognition performance.

Main Results:

• The model successfully reduced false positives caused by polysemous words through differentiated coding.
• Information sharing between different entity datasets improved the performance of each subtask.
• Achieved superior results compared to state-of-the-art methods on four typical training sets, with the best F1 values.

Conclusions:

• The proposed multi-task learning framework effectively addresses polysemy and data limitations in BioNER.
• This approach enhances the accuracy of biomedical entity recognition and offers a promising direction for future research.