BENTO

BENTO manages workflow composition and execution for clinical natural language processing to enable reproducible development and application of NLP pipelines on medical notes.

Key Features:

• Integration with CodaLab: Leverages the CodaLab platform to run and share computational workflows for clinical NLP.
• Pre-trained clinical NLP tools: Provides pre-trained models developed on datasets of medical notes with expert annotations for clinical text processing.
• Custom tool integration: Supports incorporation of user-defined NLP models and tools into composed workflows for tailored analyses.
• Flexibility and adaptability: Architecture allows adaptation of workflows and tools to domains beyond the initial clinical use cases.

Scientific Applications:

• Automated clinical text analysis: Perform tokenization and named entity recognition on medical notes.
• Reproducible workflow research: Build, share, and iterate reproducible clinical NLP pipelines.
• Scalable model application: Apply pre-trained or custom models to new datasets for large-scale clinical data analysis.

Methodology:

Built on CodaLab, BENTO executes composed workflows of pre-trained and custom NLP tools developed from annotated medical notes and supports automated tokenization and named entity recognition.