Match-Unity

Code for Long-form Text Matching with Knowledge Complementarity

Requirement

To run the code successfully, you will need (just install the most recent version):

Pytorch

Graph-tool

Jionlp

Pretrain Models

The pretrain Longformer model is from https://github.com/ValkyriaLenneth/Longformer_ZH

and The pretrain Bert model is from 'hugging face' bert-base.

How to use

1.Start by downloading the original datasets CNSE and CNSS from the repository: https://github.com/BangLiu/ArticlePairMatching.

2.Utilize data_process.py to generate JSON files and employ data_clean.py for text refinement.

3.Enhance your dataset using sup_data_advancing.py, which employs TopicTexRank to create augmented data.

4.Train an interactive matching model for long-form titles using bert_title.py.

5.Employ sentence_longformer.py to train a representational matching model tailored for lengthy contents.

6.Combine the results from steps 4 and 5 using bert_longformer.py to establish the Match-Unity model.