What Makes a Good Story? Designing Composite Rewards for Visual Storytelling
Abstract
Previous storytelling approaches mostly focused on optimizing traditional metrics such as BLEU, ROUGE and CIDEr. In this paper, we re-examine this problem from a different angle, by looking deep into what defines a natural and topically-coherent story. To this end, we propose three assessment criteria: relevance, coherence and expressiveness, which we observe through empirical analysis could constitute a “high-quality” story to the human eye. We further propose a reinforcement learning framework, ReCo-RL, with reward functions designed to capture the essence of these quality criteria. Experiments on the Visual Storytelling Dataset (VIST) with both automatic and human evaluation demonstrate that our ReCo-RL model achieves better performance than state-of-the-art baselines on both traditional metrics and the proposed new criteria.
Cite
Text
Hu et al. "What Makes a Good Story? Designing Composite Rewards for Visual Storytelling." AAAI Conference on Artificial Intelligence, 2020. doi:10.1609/AAAI.V34I05.6305Markdown
[Hu et al. "What Makes a Good Story? Designing Composite Rewards for Visual Storytelling." AAAI Conference on Artificial Intelligence, 2020.](https://mlanthology.org/aaai/2020/hu2020aaai-makes/) doi:10.1609/AAAI.V34I05.6305BibTeX
@inproceedings{hu2020aaai-makes,
title = {{What Makes a Good Story? Designing Composite Rewards for Visual Storytelling}},
author = {Hu, Junjie and Cheng, Yu and Gan, Zhe and Liu, Jingjing and Gao, Jianfeng and Neubig, Graham},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2020},
pages = {7969-7976},
doi = {10.1609/AAAI.V34I05.6305},
url = {https://mlanthology.org/aaai/2020/hu2020aaai-makes/}
}