RL for Question Generation

Exploring Question-Specific Rewards for Generating Deep Questions

This repository contains codes and models for the paper: Exploring Question-Specific Rewards for Generating Deep Questions (COLING 2020 oral). Below is the framework of our proposed model.

Requirements

Environment

pytorch 1.4.0
nltk 3.4.4
numpy 1.18.1
tqdm 4.32.2

Data Preprocessing & Models

• download here

Training

• run scripts/train_example.sh to train the ensemble model which utilizes all three rewards.

Translating

• run scripts/translate_example.sh to get the prediction on the validation dataset.

Evaluating

We take use of the Evaluation codes for MS COCO caption generation for evaluation on automatic metrics.

• To install pycocoevalcap and the pycocotools dependency, run:

pip install git+https://github.com/salaniz/pycocoevalcap

• To evaluate the results in the translated file, e.g. prediction.txt, run:

python evaluate_metrics.py prediction.txt

Citation

    @inproceedings{xie-etal-2020-RLQG,
      title = {Exploring Question-Specific Rewards for Generating Deep Questions},
      author = {Xie, Yuxi and Pan, Liangming and Wang, Dongzhe and Kan, Min-Yen and Feng, Yansong},
      booktitle = {The 28th International Conference on Computational Linguistics (COLING 2020)},
      year = {2020}
    }