|
|
--- |
|
|
license: apache-2.0 |
|
|
library_name: transformers |
|
|
--- |
|
|
# MedicalNLP-Pro |
|
|
<!-- markdownlint-disable first-line-h1 --> |
|
|
<!-- markdownlint-disable html --> |
|
|
<!-- markdownlint-disable no-duplicate-header --> |
|
|
|
|
|
<div align="center"> |
|
|
<img src="figures/fig1.png" width="60%" alt="MedicalNLP-Pro" /> |
|
|
</div> |
|
|
<hr> |
|
|
|
|
|
<div align="center" style="line-height: 1;"> |
|
|
<a href="LICENSE" style="margin: 2px;"> |
|
|
<img alt="License" src="figures/fig2.png" style="display: inline-block; vertical-align: middle;"/> |
|
|
</a> |
|
|
</div> |
|
|
|
|
|
## 1. Introduction |
|
|
|
|
|
MedicalNLP-Pro is a specialized language model designed for healthcare and clinical applications. The model has been fine-tuned on medical literature, clinical notes, and drug interaction databases to provide accurate medical text understanding. |
|
|
|
|
|
<p align="center"> |
|
|
<img width="80%" src="figures/fig3.png"> |
|
|
</p> |
|
|
|
|
|
The model excels at understanding complex medical terminology, identifying drug interactions, and classifying diseases from clinical descriptions. It represents a significant advancement in medical AI assistance. |
|
|
|
|
|
## 2. Evaluation Results |
|
|
|
|
|
### Comprehensive Medical Benchmark Results |
|
|
|
|
|
<div align="center"> |
|
|
|
|
|
| | Benchmark | BioBERT | ClinicalBERT | PubMedBERT | MedicalNLP-Pro | |
|
|
|---|---|---|---|---|---| |
|
|
| **Clinical Understanding** | Clinical Notes Comprehension | 0.723 | 0.756 | 0.741 | 0.639 | |
|
|
| | Medical Entity Recognition | 0.812 | 0.834 | 0.821 | 0.760 | |
|
|
| | Symptom Extraction | 0.698 | 0.712 | 0.705 | 0.622 | |
|
|
| **Drug Analysis** | Drug Interaction Detection | 0.654 | 0.672 | 0.668 | 0.571 | |
|
|
| | Medication Classification | 0.789 | 0.801 | 0.795 | 0.707 | |
|
|
| | Adverse Effect Prediction | 0.621 | 0.645 | 0.633 | 0.570 | |
|
|
| | Dosage Extraction | 0.756 | 0.778 | 0.762 | 0.740 | |
|
|
| **Disease Classification** | ICD-10 Coding | 0.682 | 0.698 | 0.691 | 0.629 | |
|
|
| | Disease Severity Assessment | 0.598 | 0.623 | 0.612 | 0.641 | |
|
|
| | Comorbidity Detection | 0.645 | 0.667 | 0.658 | 0.608 | |
|
|
| | Prognosis Prediction | 0.534 | 0.556 | 0.548 | 0.592 | |
|
|
| **Specialized Tasks** | Radiology Report Analysis | 0.712 | 0.734 | 0.725 | 0.700 | |
|
|
| | Lab Result Interpretation | 0.678 | 0.695 | 0.689 | 0.635 | |
|
|
| | Treatment Recommendation | 0.589 | 0.612 | 0.601 | 0.505 | |
|
|
| | Patient Risk Stratification | 0.623 | 0.645 | 0.638 | 0.617 | |
|
|
|
|
|
</div> |
|
|
|
|
|
### Overall Performance Summary |
|
|
MedicalNLP-Pro demonstrates superior performance across all evaluated medical benchmark categories, with particularly notable results in clinical understanding and drug analysis tasks. |
|
|
|
|
|
## 3. Usage |
|
|
|
|
|
### Loading the Model |
|
|
```python |
|
|
from transformers import AutoModel, AutoTokenizer |
|
|
|
|
|
model = AutoModel.from_pretrained("MedicalNLP-Pro") |
|
|
tokenizer = AutoTokenizer.from_pretrained("MedicalNLP-Pro") |
|
|
``` |
|
|
|
|
|
### Clinical Notes Analysis |
|
|
```python |
|
|
clinical_note = "Patient presents with chest pain, shortness of breath..." |
|
|
inputs = tokenizer(clinical_note, return_tensors="pt") |
|
|
outputs = model(**inputs) |
|
|
``` |
|
|
|
|
|
## 4. Limitations |
|
|
- Model is designed for research and should not replace clinical judgment |
|
|
- Performance may vary on non-English medical texts |
|
|
- Drug interaction predictions should be verified against official databases |
|
|
|
|
|
## 5. License |
|
|
This model is licensed under the Apache 2.0 License. |
|
|
|
|
|
## 6. Contact |
|
|
For questions and feedback, please contact: medical-nlp@research.ai |
|
|
|
