KEHNN

Source code of our AAAI 2016 submitted paper, Knowledge Enhanced Hybrid Neural Network for Text Matching. In general, this is an experiment level source code, consisting of "hard code" and lacking of "code comment". "KEHNN.py" is the core file in this project

Input format

question knowledge \t answer knowledge \t lable \t question \answer

Usage

First, you should use process_ubuntu.py to parse your input file, and our code will generate a Bin file for "KEHNN.py".