Introduction

• This is a competition using doctor-patient dialogue finding privacy sensitive span.

Model Structure

• Input doctor-patient dialogue (split by speaker) through language model.
• Each word embedding go through 2 different feed-dorward network get BIO type and privacy type.
  • BIO type use for checking what span need to be output.
  • Privacy type use for checking what type of privacy sensitive to be output.

Usage

• All file have argument parser, please use python3 FILENAME -h to get more.

Acknowledgments

• Thank @eric88525 doing data preprocessing (preprocess_to_json.py); @HongYun0901 pretraining the lenguage model.