Answerability-Metric

This repository contains the script to compute the questions based on the Answerability aspect. Towards a Better Metric for Evaluating Question Generation Systems - Preksha Nema and Mitesh Khapra ( EMNLP, 2018)

Implementation

The script is a modified version of coco-caption.

Requirements

• python setup.py install
• Java
• pickle
• NIST (Optional)
• METEOR (Optional)

Data Format

• The reference file and the hypothesis file should contain one question per line.
• NIST, METEOR scores should be precomputed using the given script, and should contain score for (ref_ques, hyp_ques) per line, corresponding to the ref_file, hyp_file.

How to run the script

Run:

python answerability_score.py --data_type squad --ref_file examples/references.txt --hyp_file examples/hypotheses.txt --ner_weight 0.6 --qt_weight 0.2 --re_weight 0.1 --delta 0.7 --ngram_metric Bleu_3

• ngram_metric varies from [Bleu_1, Bleu_2, Bleu_3, Bleu_4, ROUGE_L, NIST, METEOR]

Note that the weights given to the stop_words is computed as (1 - ner_weight - qt_weight - re_weight) and the weight for the ngram metric is computed using (1 - delta)

Testing

Run:

pip install -e '.[dev]'
pytest