Controllable Text Simplification with Explicit Paraphrasing

This repository contains the code and resources from the following paper. Our approach simplifies the given complex sentence in three steps:

1. Generate candidates for an input sentence using DisSim and neural sentence splitter. DisSim is a rule-based approach proposed by Nikluas et al. 2019 that uses 35 syntactic rules to split a sentence.

2. Rank the candidates that have undergone splitting and deletion based on the quality of simplification.

3. Pass the best ranked candidate to the paraphrase generation Transformer model.

Candidate Generation:

First, you need to install the DiscourseSimplification code. We use the same code from this repo.

cd DiscourseSimplification
mvn clean install -DskipTests

To generate the candidates, you can use the following command:

python3 generate_candidates.py --input <input filename> --output <candidate filename>

Candidate Ranking:

To rank the candidates generated in the previous step, you can use the following command:

python3 ranking/main.py --input <input filename> --candidates <candidate filename> --output <best ranked candidate filename>

Paraphrase Generation:

Coming Soon.

Citation

Please cite if you use the above resources for your research

@InProceedings{NAACL-2021-Maddela,
  author = 	"Maddela, Mounica and Alva-Manchego, Fernando and Xu, Wei",
  title = 	"Controllable Text Simplification with Explicit Paraphrasing",
  booktitle = 	"Proceedings of the North American Association for Computational Linguistics (NAACL)",
  year = 	"2021",
}