klauskades/n2c2-challenge-2019

n2c2 Challenge 2019

This repository contains our contribution to the 2019 n2c2 challenge for track 1 ("n2c2/OHNLP Track on Clinical Semantic Textual Similarity"). The code in this repository can be used to reproduce the results in our paper "Extending BERT for Clinical Semantic Textual Similarity".

How to Use the Code

• Make sure you have Docker installed on your system.
• Navigate to the docker directory and build the image: docker build --tag n2c2 .
• Set the following environment variables:
  • NLP_MODELS_PATH: path to the directory where you downloaded the models. This directory should have the following structure (for the BERT model see https://github.com/EmilyAlsentzer/clinicalBERT and the InferSent (and dependencies) models can be found at https://github.com/facebookresearch/InferSent):

  models
  └─ pretrained
     ├── document_embeddings
     └── word_embeddings
         ├── bert_models
         │   └── biobert_pretrain_output_all_notes_150000
         │       ├── bert_config.json
         │       ├── bert_model.ckpt.data-00000-of-00001
         │       ├── bert_model.ckpt.index
         │       ├── bert_model.ckpt.meta
         │       ├── graph.pbtxt
         │       ├── pytorch_model.bin
         │       └── vocab.txt
         ├── crawl-300d-2M
         │   ├── crawl-300d-2M.vec
         │   └── model_info.md
         ├── glove.840B.300d
         │   ├── glove.840B.300d.txt
         │   └── model_info.md
         ├── infersent1
         │   ├── infersent1.pkl
         │   └── model_info.md
         └── infersent2
             ├── infersent2.pkl
             └── model_info.md
  

  • NLP_RAW_DATA: path to the directory which ontains the raw challenge data with the following directory structure:

  challenge_data
  └── n2c2
      ├── clinicalSTS2019.test.gs.sim.txt
      ├── clinicalSTS2019.test.txt
      └── clinicalSTS2019.train.txt
  

  • NLP_EXPERIMENT_PATH: path to a directory which is used to store the results from the model.
• Run the container (also inside the docker folder): docker run --gpus all -it --rm -v ${PWD}/..:/workspace -v $NLP_MODELS_PATH:/mtc/models -v $NLP_RAW_DATA:/mtc/raw -v $NLP_EXPERIMENT_PATH:/mtc/experiment --name n2c2 n2c2
• You can now execute the file generate_results.py inside the container.
• If everything was successful, there should now be a subfolder named submission_generation which contains the resulting scores which we report in our paper.

Copyright

Copyright © German Cancer Research Center (DKFZ), Division of Medical Image Computing (MIC). Please make sure that your usage of this code is in compliance with the code license.