Skip to content

Latest commit

 

History

History
15 lines (8 loc) · 1.02 KB

README.md

File metadata and controls

15 lines (8 loc) · 1.02 KB
Raw

A Lightweight Method to Generate Unanswerable Questions in English

This repository contains code and data for our paper.

Code

Code for generating unanswerable questions using entity and antonym swaps, and for filtering them (random filtering as well as GPT2-perplexity-based filtering) are located in the generating swaps folder. The finetuning for QA folder contains the code we used to train BERT, RoBERTa and ALBERT models for question answering. As input and output, this code uses JSON files in the SQuAD 2.0 format.

Data

For reproducibility, we provide all of our generated and filtered unanswerable questions in the augmented data folder. Along with our finetuning for QA code, this data can be used to reproduce our reported numbers for our methods as seen in the body of the paper.

Human evaluation

We also release the data and annotations (from 3 annotators) from our human evaluation. The annotator instructions are available in Appendix C of the paper.