Gary-code / MOAG

The basic code for ACM MM 2021 paper "Multiple Objects-Aware Visual Question Generation" and Neural Network 2023 paper "Visual Question Generation for Explicit Questioning Purposes Based on Target Objects"

MOAG

The reproducing code for ACM MM 2021 paper titled "Multiple Objects-Aware Visual Question Generation." [paper]

Overview

Installation

• PyTorch = 1.12

Run MOAG

python train.py

Reference

@inproceedings{moag,
author = {Xie, Jiayuan and Cai, Yi and Huang, Qingbao and Wang, Tao},
title = {Multiple Objects-Aware Visual Question Generation},
year = {2021},
booktitle = {Proceedings of the 29th ACM International Conference on Multimedia},
pages = {4546–4554},
}

News-CcQG

Our Content-controlled Question Generation (CcQG) model is extension model of MOAG, details can be found in Neural Network 2023 paper “Visual Question Generation for Explicit Questioning Purposes Based on Target Objects” [pdf]

@article{ccqg,
  title={Visual question generation for explicit questioning purposes based on target objects},
  author={Xie, Jiayuan and Chen, Jiali and Fang, Wenhao and Cai, Yi and Li, Qing},
  journal={Neural Networks},
  volume={167},
  pages={638--647},
  year={2023},
  publisher={Elsevier}
}