Named Entity Recognition for Biomedical Text
Develop an accurate model for predicting named entities in unlabelled biomedical text using the comprehensive NCBI disease corpus. This project empowers the biomedical NLP community with a valuable tool for automatic entity extraction, advancing research and supporting clinical decision-making.
Natural Language Processing (NLP)
Named Entity Recognition (NER)
The objective is to develop models capable of producing accurate and coherent summaries based on the article's title and text. Leveraging BERT, the models will learn intricate patterns and relationships within the training data, enabling them to automatically generate summaries that capture the main points and key information from the articles.The training data consists of a list of dictionaries in JSON format, where each element represents a distinct article with "title," "text," and "summary" keys.
The models will be fine-tuned using BERT's pre-trained weights and optimised through attention mechanisms, enabling them to effectively extract salient information for summary generation.
This project yields powerful tools for automatically generating extractive summaries for text articles. These models can be applied to various practical applications, including information retrieval systems, content summarization for news articles, blogs, or research papers, and aiding users in quickly comprehending lengthy texts.
Tools & Technology
Programming Language: Python