An official website of the United States government
Showing 4 Results
Showing 1-4 of 4
Model & Software
To view details of each card, click 
icon
Project:
MOSSAIC
Description:
Offers a hierarchical self-attention network for information extraction from long (more than 400 words) cancer pathology reports.
DESCRIPTION:
Offers a hierarchical self-attention network for information extraction from long (more than 400 words) cancer pathology reports.
IMPACT: Allows automatic information extraction from free-form pathology report texts. More accurate than MT-CNN.
PRIMARY PUBLICATION: Limitations of Transformers on Clinical Text Classification
INPUT DATA FORMAT: Unspecified
LEVEL OF DOCUMENTATION: Minimal
AVAILABLE ON GITHUB
Project:
MOSSAIC
Description:
Offers a convolutional neural network for natural language processing and information extraction from free-form texts.
DESCRIPTION:
Offers a convolutional neural network for natural language processing and information extraction from free-form texts.
IMPACT: Allows automatic information extraction from free-form pathology report texts. Faster than HiSAN.
PRIMARY PUBLICATION: Deep Active Learning for Classifying Cancer Pathology Reports
INPUT DATA TYPE: Tokenized Text
INPUT DATA FORMAT: Unspecified
LEVEL OF DOCUMENTATION: Minimal
AVAILABLE ON GITHUB
Project:
MOSSAIC
Description:
Builds a network to automatically extract information from biomedical text.
DESCRIPTION:
Builds a network to automatically extract information from biomedical text.
IMPACT: Allows automatic extraction of cancer-relevant information from free-text pathology reports.
PRIMARY PUBLICATION: Multi-task Deep Neural Networks for Automated Extraction of Primary Site and Laterality Information from Cancer Pathology Reports
INPUT DATA FORMAT: Tabular
LEVEL OF DOCUMENTATION: Minimal
AVAILABLE ON GITHUB
Project:
MOSSAIC
Description:
Generates synthetic biomedical text of desired clinical context.
DESCRIPTION:
Generates synthetic biomedical text of desired clinical context.
IMPACT: Creates examples of unstructured text with a specific label from a given corpus that can be used for training machine learning or deep learning models on clinical text. Labeled data is quite challenging to come by, specifically for patient data, for deep text comprehension applications.
INPUT DATA FORMAT: Unspecified
LEVEL OF DOCUMENTATION: Minimal
AVAILABLE ON GITHUB