A Summarize-then-Search Method for Long Video Question Answering: Limitations & References

Table of Links

• Abstract and Intro
• Method
• Experiments
• Related Work
• Conclusion
• Limitations and References
• A. Experiment Details
• B. Prompt Samples

6. Limitations

Our study has some limitations, including:

1. We experiment with only videos with English subtitles. However, our method can be extended to include multi-lingual contexts given a strong multilingual language model.

2. The computation and memory requirement of our method is substantial due to its heavy reliance on the large language model, GPT-3.

3. We evaluate Long Story Short with only a single instance of LLM (GPT-3).

Potential Risk. Summarizing the long video context with GPT-3 carries on ethical risks related to the open-ended nature of the language model. GPT-3 may (a) hallucinate fake facts about the content, (b) generate toxic utterances, or (c) implicitly embed social biases into the summary and the answer likelihoods.

References

[1] Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in neural information processing systems, 33:1877–1901, 2020.

[2] Seongho Choi, Kyoung-Woon On, Yu-Jung Heo, Ahjeong Seo, Youwon Jang, Seungchan Lee, Minsu Lee, and Byoung-Tak Zhang. DramaQA: character-centered video story understanding with hierarchical qa. arXiv preprint arXiv:2005.03356, 2020.

[3] Seongho Choi, Kyoung-Woon On, Yu-Jung Heo, Ahjeong Seo, Youwon Jang, Minsu Lee, and Byoung-Tak Zhang. Dramaqa: Character-centered video story understanding with hierarchical qa. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 1166–1174, 2021.

[4] Chenyou Fan, Xiaofan Zhang, Shu Zhang, Wensheng Wang, Chi Zhang, and Heng Huang. Heterogeneous memory enhanced multimodal attention model for video question answering. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1999–2007, 2019.

[5] Tsu-Jui Fu, Linjie Li, Zhe Gan, Kevin Lin, William Yang Wang, Lijuan Wang, and Zicheng Liu. Violet: End-to-end video-language transformers with masked visual-token modeling. arXiv preprint arXiv:2111.12681, 2021.

[6] Jiyang Gao, Runzhou Ge, Kan Chen, and Ram Nevatia. Motion-appearance co-memory networks for video question answering. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 6576–6585, 2018.

[7] Philip John Gorinski and Mirella Lapata. Movie script summarization as graph-based scene extraction. In NAACL, 2015.

[8] Pengcheng He, Baolin Peng, Liyang Lu, Songhe Wang, Jie Mei, Yang Liu, Ruochen Xu, Hany Hassan Awadalla, Yu Shi, Chenguang Zhu, Wayne Xiong, Michael Zeng, Jianfeng Gao, and Xuedong Huang. Z-code++: A pre-trained language model optimized for abstractive summarization. ArXiv, abs/2208.09770, 2022.

[9] Yunseok Jang, Yale Song, Youngjae Yu, Youngjin Kim, and Gunhee Kim. Tgif-qa: Toward spatio-temporal reasoning in visual question answering. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2758–2766, 2017.

[10] Bhavan Jasani, Rohit Girdhar, and Deva Ramanan. Are we asking the right questions in movieqa? In Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, pages 0–0, 2019.

[11] Junyeong Kim, Minuk Ma, Kyungsu Kim, Sungjin Kim, and Chang D Yoo. Progressive attention memory network for movie story question answering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8337–8346, 2019.

[12] Junyeong Kim, Minuk Ma, Kyungsu Kim, Sungjin Kim, and Chang D Yoo. Progressive attention memory network for movie story question answering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8337–8346, 2019.

[13] Kyung-Min Kim, Min-Oh Heo, Seong-Ho Choi, and Byoung-Tak Zhang. Deepstory: video story qa by deep embedded memory networks. In Proceedings of the 26th International Joint Conference on Artificial Intelligence, pages 2016–2022, 2017.

[14] Seonhoon Kim, Seohyeong Jeong, Eunbyul Kim, Inho Kang, and Nojun Kwak. Selfsupervised pre-training and contrastive representation learning for multiple-choice video qa. In AAAI, 2021.

[15] Myungji Lee, Hong-Seok Kwon, Jaehun Shin, WonKee Lee, Baikjin Jung, and JongHyeok Lee. Transformer-based screenplay summarization using augmented learning representation with dialogue information. In NUSE, 2021.

[16] Jie Lei, Licheng Yu, Mohit Bansal, and Tamara L Berg. Tvqa: Localized, compositional video question answering. In EMNLP, 2018.

[17] Jie Lei, Licheng Yu, Tamara L Berg, and Mohit Bansal. Tvqa+: Spatio-temporal grounding for video question answering. In Tech Report, arXiv, 2019.

[18] Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. Blip: Bootstrapping languageimage pre-training for unified vision-language understanding and generation. In ICML, 2022.

[19] Chin-Yew Lin. ROUGE: A package for automatic evaluation of summaries. In Text Summarization Branches Out, pages 74–81, Barcelona, Spain, July 2004. Association for Computational Linguistics. URL https://aclanthology.org/W04-1013.

[20] Chao-Ning Liu, Ding-Jie Chen, Hwann-Tzong Chen, and Tyng-Luh Liu. A2a: Attention to attention reasoning for movie question answering. In Computer Vision–ACCV 2018: 14th Asian Conference on Computer Vision, Perth, Australia, December 2–6, 2018, Revised Selected Papers, Part VI 14, pages 404–419. Springer, 2019.

[21] Fei Liu, Jing Liu, Xinxin Zhu, Richang Hong, and Hanqing Lu. Dual hierarchical temporal convolutional network with qa-aware dynamic normalization for video story question answering. In Proceedings of the 28th ACM International Conference on Multimedia, pages 4253–4261, 2020.

[22] Seil Na, Sangho Lee, Jisung Kim, and Gunhee Kim. A read-write memory network for movie story understanding. In Proceedings of the IEEE International Conference on Computer Vision, pages 677–685, 2017.

[23] Pinelopi Papalampidi, Frank Keller, and Mirella Lapata. Movie plot analysis via turning point identification. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), November 2019.

[24] Pinelopi Papalampidi, Frank Keller, Lea Frermann, and Mirella Lapata. Screenplay summarization using latent narrative structure. In Annual Meeting of the Association for Computational Linguistics, 2020.

[25] Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In International Conference on Machine Learning, pages 8748–8763. PMLR, 2021.

[26] Anna Rohrbach, Atousa Torabi, Marcus Rohrbach, Niket Tandon, Christopher Pal, Hugo Larochelle, Aaron Courville, and Bernt Schiele. Movie Description. IJCV, 2017.

[27] Makarand Tapaswi, Yukun Zhu, Rainer Stiefelhagen, Antonio Torralba, Raquel Urtasun, and Sanja Fidler. Movieqa: Understanding stories in movies through question-answering. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4631–4640, 2016.

[28] Bo Wu, Shoubin Yu, Zhenfang Chen, Joshua B Tenenbaum, and Chuang Gan. Star: A benchmark for situated reasoning in real-world videos. In Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2), 2021.

[29] Junbin Xiao, Xindi Shang, Angela Yao, and Tat-Seng Chua. Next-qa: Next phase of question-answering to explaining temporal actions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9777–9786, 2021.

[30] Dejing Xu, Zhou Zhao, Jun Xiao, Fei Wu, Hanwang Zhang, Xiangnan He, and Yueting Zhuang. Video question answering via gradually refined attention over appearance and motion. In Proceedings of the 25th ACM international conference on Multimedia, pages 1645–1653, 2017.

[31] Antoine Yang, Antoine Miech, Josef Sivic, Ivan Laptev, and Cordelia Schmid. Just ask: Learning to answer questions from millions of narrated videos. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 1686–1697, 2021.

[32] Zhengyuan Yang, Zhe Gan, Jianfeng Wang, Xiaowei Hu, Yumao Lu, Zicheng Liu, and Lijuan Wang. An empirical study of gpt-3 for few-shot knowledge-based vqa. arXiv preprint arXiv:2109.05014, 2021.

[33] Rowan Zellers, Ximing Lu, Jack Hessel, Youngjae Yu, Jae Sung Park, Jize Cao, Ali Farhadi, and Yejin Choi. Merlot: Multimodal neural script knowledge models. In M. Ranzato, A. Beygelzimer, Y. Dauphin, P.S. Liang, and J. Wortman Vaughan, editors, Advances in Neural Information Processing Systems, volume 34, pages 23634–23651. Curran Associates, Inc., 2021. URL https://proceedings.neurips.cc/paper/ 2021/file/c6d4eb15f1e84a36eff58eca3627c82e-Paper.pdf.

[34] Rowan Zellers, Jiasen Lu, Ximing Lu, Youngjae Yu, Yanpeng Zhao, Mohammadreza Salehi, Aditya Kusupati, Jack Hessel, Ali Farhadi, and Yejin Choi. Merlot reserve: Neural script knowledge through vision and language and sound. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.

[35] Andy Zeng, Adrian Wong, Stefan Welker, Krzysztof Choromanski, Federico Tombari, Aveek Purohit, Michael S Ryoo, Vikas Sindhwani, Johnny Lee, Vincent Vanhoucke, et al. Socratic models: Composing zero-shot multimodal reasoning with language. 2022.

[36] Kuo-Hao Zeng, Tseng-Hung Chen, Ching-Yao Chuang, Yuan-Hong Liao, Juan Carlos Niebles, and Min Sun. Leveraging video descriptions to learn video question answering. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 31, 2017.

[37] Jingqing Zhang, Yao Zhao, Mohammad Saleh, and Peter Liu. Pegasus: Pre-training with extracted gap-sentences for abstractive summarization. In International Conference on Machine Learning, pages 11328–11339. PMLR, 2020.

[38] Zhou Zhao, Jinghao Lin, Xinghua Jiang, Deng Cai, Xiaofei He, and Yueting Zhuang. Video question answering via hierarchical dual-level attention network learning. In Proceedings of the 25th ACM international conference on Multimedia, pages 1050–1058, 2017.