@inproceedings{wu-etal-2026-knowmebenchv2,
title = "{K}now{M}e{B}ench{V}2: Evidence-Grounded Person-Centric Long-Video Understanding",
author = "Wu, Tingyu and
Cao, Guangyu and
Lan, Qizhen and
Chen, Zhisheng and
Weng, Ziyan and
Zhu, Bingkun and
Su, Miao and
Li, Chenglong and
Xie, Zhengwei and
Wang, Huacan and
Hu, Sen and
Ye, Haocun and
AN, Nan and
Liu, Zaoqu and
Chen, Ronghao",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.811/",
pages = "18893--18913",
ISBN = "979-8-89176-459-0",
abstract = "Questions about a recurring person in long videos are rarely answered from a single moment. They require evidence that is sufficient for the claim, comparisons across episodes, synthesis over recurring audiovisual patterns, and refusal when the source material does not support an answer. Existing long-video benchmarks do not isolate these requirements. We introduce KnowMeBenchV2, an evaluation-only benchmark over five public multi-episode video series, with 327 videos and 3,061 questions covering four evidence scopes: single-clip extraction, temporal and event relations, cross-episode ordering, and cross-episode persona synthesis. A separate grounded-abstention task serves as a contrastive support check: its questions are surface-similar to supported person-linked questions but are not licensed by the source evidence. The most interpretive synthesis items are human-authored and quality-audited by trained annotators with full source access. A matched human reference reaches 78.0 on cross-episode synthesis. To diagnose failures, we compare evidence routes under a fixed Qwen3.5-9B answerer and a task-matched pre-tokenization character routing budget. ASR-Retrieval reaches 2.7 on synthesis, while KnowMe reaches 64.9. In this diagnostic, the final answerer reasons over textualized evidence constructed from audiovisual sources rather than raw video tokens. The gap therefore isolates the importance of audiovisual evidence construction and cross-episode organization before answer generation."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wu-etal-2026-knowmebenchv2">
<titleInfo>
<title>KnowMeBenchV2: Evidence-Grounded Person-Centric Long-Video Understanding</title>
</titleInfo>
<name type="personal">
<namePart type="given">Tingyu</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guangyu</namePart>
<namePart type="family">Cao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qizhen</namePart>
<namePart type="family">Lan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhisheng</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ziyan</namePart>
<namePart type="family">Weng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bingkun</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Miao</namePart>
<namePart type="family">Su</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chenglong</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengwei</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Huacan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sen</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haocun</namePart>
<namePart type="family">Ye</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nan</namePart>
<namePart type="family">AN</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zaoqu</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ronghao</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Questions about a recurring person in long videos are rarely answered from a single moment. They require evidence that is sufficient for the claim, comparisons across episodes, synthesis over recurring audiovisual patterns, and refusal when the source material does not support an answer. Existing long-video benchmarks do not isolate these requirements. We introduce KnowMeBenchV2, an evaluation-only benchmark over five public multi-episode video series, with 327 videos and 3,061 questions covering four evidence scopes: single-clip extraction, temporal and event relations, cross-episode ordering, and cross-episode persona synthesis. A separate grounded-abstention task serves as a contrastive support check: its questions are surface-similar to supported person-linked questions but are not licensed by the source evidence. The most interpretive synthesis items are human-authored and quality-audited by trained annotators with full source access. A matched human reference reaches 78.0 on cross-episode synthesis. To diagnose failures, we compare evidence routes under a fixed Qwen3.5-9B answerer and a task-matched pre-tokenization character routing budget. ASR-Retrieval reaches 2.7 on synthesis, while KnowMe reaches 64.9. In this diagnostic, the final answerer reasons over textualized evidence constructed from audiovisual sources rather than raw video tokens. The gap therefore isolates the importance of audiovisual evidence construction and cross-episode organization before answer generation.</abstract>
<identifier type="citekey">wu-etal-2026-knowmebenchv2</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.811/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>18893</start>
<end>18913</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T KnowMeBenchV2: Evidence-Grounded Person-Centric Long-Video Understanding
%A Wu, Tingyu
%A Cao, Guangyu
%A Lan, Qizhen
%A Chen, Zhisheng
%A Weng, Ziyan
%A Zhu, Bingkun
%A Su, Miao
%A Li, Chenglong
%A Xie, Zhengwei
%A Wang, Huacan
%A Hu, Sen
%A Ye, Haocun
%A AN, Nan
%A Liu, Zaoqu
%A Chen, Ronghao
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F wu-etal-2026-knowmebenchv2
%X Questions about a recurring person in long videos are rarely answered from a single moment. They require evidence that is sufficient for the claim, comparisons across episodes, synthesis over recurring audiovisual patterns, and refusal when the source material does not support an answer. Existing long-video benchmarks do not isolate these requirements. We introduce KnowMeBenchV2, an evaluation-only benchmark over five public multi-episode video series, with 327 videos and 3,061 questions covering four evidence scopes: single-clip extraction, temporal and event relations, cross-episode ordering, and cross-episode persona synthesis. A separate grounded-abstention task serves as a contrastive support check: its questions are surface-similar to supported person-linked questions but are not licensed by the source evidence. The most interpretive synthesis items are human-authored and quality-audited by trained annotators with full source access. A matched human reference reaches 78.0 on cross-episode synthesis. To diagnose failures, we compare evidence routes under a fixed Qwen3.5-9B answerer and a task-matched pre-tokenization character routing budget. ASR-Retrieval reaches 2.7 on synthesis, while KnowMe reaches 64.9. In this diagnostic, the final answerer reasons over textualized evidence constructed from audiovisual sources rather than raw video tokens. The gap therefore isolates the importance of audiovisual evidence construction and cross-episode organization before answer generation.
%U https://aclanthology.org/2026.emnlp-main.811/
%P 18893-18913
Markdown (Informal)
[KnowMeBenchV2: Evidence-Grounded Person-Centric Long-Video Understanding](https://aclanthology.org/2026.emnlp-main.811/) (Wu et al., EMNLP 2026)
ACL
- Tingyu Wu, Guangyu Cao, Qizhen Lan, Zhisheng Chen, Ziyan Weng, Bingkun Zhu, Miao Su, Chenglong Li, Zhengwei Xie, Huacan Wang, Sen Hu, Haocun Ye, Nan AN, Zaoqu Liu, and Ronghao Chen. 2026. KnowMeBenchV2: Evidence-Grounded Person-Centric Long-Video Understanding. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 18893–18913, Budapest, Hungary. Association for Computational Linguistics.