CogStream: Context-guided Streaming Video Question Answering
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (16) · 13332-13341
DOI: 10.1609/aaai.v40i16.38336 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Zicheng Zhao und 5 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref).
Abstract (Originalsprache)
Despite advancements in Video Large Language Models (Vid-LLMs) improving multimodal understanding, challenges persist in streaming video reasoning due to its reliance on contextual information. Existing paradigms feed all available historical contextual information into Vid-LLMs, resulting in a significant computational burden for visual data processing. Furthermore, the inclusion of irrelevant context distracts models from key details. This paper introduces a challenging task called Context-guided Streaming Video Reasoning (CogStream), which simulates real-world streaming video scenarios, requiring models to identify the most relevant historical contextual information to deduce answers for questions about the current stream. To support CogStream, we present a densely annotated dataset featuring extensive and hierarchical question-answer pairs, generated by a semi-automatic pipeline. Additionally, we present CogReasoner as a baseline model. It effectively tackles this task by leveraging visual stream compression and historical dialogue retrieval. Extensive experiments prove the effectiveness of this method.
Zitieren
APA
Zhao, Z., Wang, K., Li, S., Qian, R., Lin, W., & Liu, H. (2026). CogStream: Context-guided Streaming Video Question Answering. Proceedings of the AAAI Conference on Artificial Intelligence, 40(16), 13332-13341. https://doi.org/10.1609/aaai.v40i16.38336
Vancouver
Zhao Z, Wang K, Li S, Qian R, Lin W, Liu H. CogStream: Context-guided Streaming Video Question Answering. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(16):13332-13341. doi:10.1609/aaai.v40i16.38336
BibTeX
@article{Zhao2026,
author = {Zhao, Zicheng and Wang, Kangyu and Li, Shijie and Qian, Rui and Lin, Weiyao and Liu, Huabin},
title = {CogStream: Context-guided Streaming Video Question Answering},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {16},
pages = {13332--13341},
doi = {10.1609/aaai.v40i16.38336}
}
EUVIMED