microsoft/BioGPT
BioGPT
BioGPT is a biomedical large language model and domain-specific transformer designed for processing and creating specialized medical text. It functions as a generative tool and knowledge extraction engine trained on large-scale scientific literature to produce human-like scientific prose and factual responses to queries.
The project provides specialized capabilities for biomedical named entity recognition and the extraction of complex relations from unstructured medical corpora. It is designed to identify and classify biological entities through data mining and relation extraction to support downstream medical analysis.
The system covers a broad range of scientific text processing tasks, including biomedical document classification and question answering. It also supports domain-specific model fine-tuning and text feature extraction to adapt the pre-trained architecture for niche scientific tasks.
Features
- Biomedical Relation Extractors - Identifies relationships between biomedical entities such as chemicals, diseases, and drugs from text.
- Biomedical Fine-Tuning Pipelines - Adapts a pre-trained biomedical language model to specific relation extraction tasks using domain-specific data.
- Medical Relationship Extraction - Extracts structured medical relations from biomedical text using a fine-tuned inference model.
- Biomedical Entity Recognizers - Provides a specialized model for identifying and classifying biological entities in medical text.
- Biomedical Pre-Trained Transformers - Ships a transformer model pre-trained on biomedical literature for domain-specific text understanding.
- Relation Extraction - Extracts structured relationships between biomedical entities by processing text through a fine-tuned model.
- Biomedical Text Generators - Generates human-like scientific prose and factual responses based on medical domain knowledge.
- Literature Mining Pipelines - Extracts structured information from biomedical literature using a specialized language model.
- Fine-Tuning Pipelines - Provides workflows for adapting the pre-trained model to specific biomedical tasks via supervised training.
- Task-Specific Fine-Tuning Pipelines - Provides pipelines for adapting the base model to specific downstream tasks like classification and QA.
- Fine-Tuned Model Evaluators - Runs inference on test datasets and computes performance metrics to assess model accuracy.
- Biomedical Document Classification - Assigns biomedical documents to predefined categories using a fine-tuned transformer model.
- Model Evaluation Frameworks - Includes utilities for running model inference and validation on standard biomedical datasets.
- Model Generalization - Runs inference on held-out test sets and computes accuracy metrics to measure generalization.
- Biomedical - Answers questions from biomedical research by fine-tuning on PubMedQA datasets.
- Question Answering - Answers medical questions by fine-tuning the model on curated biomedical question-answer datasets.
- Question Answering Benchmarks - Measures model performance on question-answering tasks using test sets with ground truth labels.
- Test Set Inference Evaluators - Runs trained models on held-out data and computes performance metrics to measure accuracy.
- Domain Specific Bots - Focuses on biomedical research and entity extraction.
- GPT-2 models - Listed in the “GPT-2 models” section of the Awesome Bioie awesome list.
Star history
How this analysis was created: This summary and feature list were written by an AI model that read the project's README and public documentation pages. Each feature links to the documentation it came from; stars, license and language come straight from the GitHub API. The model does not read the source code, and the analysis is refreshed when the project is re-analysed. Learn more on our About page.
