CDGP: Automatic Cloze Distractor Generation based on Pre-trained Language Model

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chiang, Shang-Hsuan, Wang, Ssu-Cheng, Fan, Yao-Chung
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917614711734272
author Chiang, Shang-Hsuan
Wang, Ssu-Cheng
Fan, Yao-Chung
author_facet Chiang, Shang-Hsuan
Wang, Ssu-Cheng
Fan, Yao-Chung
contents Manually designing cloze test consumes enormous time and efforts. The major challenge lies in wrong option (distractor) selection. Having carefully-design distractors improves the effectiveness of learner ability assessment. As a result, the idea of automatically generating cloze distractor is motivated. In this paper, we investigate cloze distractor generation by exploring the employment of pre-trained language models (PLMs) as an alternative for candidate distractor generation. Experiments show that the PLM-enhanced model brings a substantial performance improvement. Our best performing model advances the state-of-the-art result from 14.94 to 34.17 (NDCG@10 score). Our code and dataset is available at https://github.com/AndyChiangSH/CDGP.
format Preprint
id arxiv_https___arxiv_org_abs_2403_10326
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CDGP: Automatic Cloze Distractor Generation based on Pre-trained Language Model
Chiang, Shang-Hsuan
Wang, Ssu-Cheng
Fan, Yao-Chung
Computation and Language
Artificial Intelligence
Machine Learning
Manually designing cloze test consumes enormous time and efforts. The major challenge lies in wrong option (distractor) selection. Having carefully-design distractors improves the effectiveness of learner ability assessment. As a result, the idea of automatically generating cloze distractor is motivated. In this paper, we investigate cloze distractor generation by exploring the employment of pre-trained language models (PLMs) as an alternative for candidate distractor generation. Experiments show that the PLM-enhanced model brings a substantial performance improvement. Our best performing model advances the state-of-the-art result from 14.94 to 34.17 (NDCG@10 score). Our code and dataset is available at https://github.com/AndyChiangSH/CDGP.
title CDGP: Automatic Cloze Distractor Generation based on Pre-trained Language Model
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2403.10326