@inproceedings{cao-etal-2026-dramasearch,
title = "{D}rama{S}earch: Benchmarking Multimodal Searching Agents in Ultra-long {TV} Dramas",
author = "Cao, Yunning and
Ge, Jiannan and
Liu, Chuanbin and
Xie, Lingxi and
Huo, Xinyue and
Chen, Pengfei and
Duan, Kaiwen and
Xie, Hongtao and
Zhang, Yongdong and
Tian, Qi",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1313/",
pages = "27798--27826",
ISBN = "979-8-89176-462-0",
abstract = "Searching agents have achieved significant progress in NLP, yet their multimodal counterparts are constrained by a lack of suitable evaluation benchmarks requiring reasoning over continuous video, audio, and images. We propose long-form TV dramas as a natural solution and introduce DramaSearch, encompassing 3,600 riddles across 9 dramas totaling 239 hours. Our dataset is generated via a graph-constrained pipeline that processes enriched captions into structured events and multi-granularity evidence views. By employing character-to-view and entity-to-view indexing, we ensure controllable clue disclosure and strictly unique answers. The riddles are categorized into multi-clue, multi-hop, and compositional types, requiring agents to aggregate isolated clues and navigate complex dependency chains. Experiments demonstrate that a Qwen3-8B agent, trained on a held-out set using SFT and RL, achieves an accuracy of 48.3{\%}. This result establishes a strong baseline that performs favorably against zero-shot GPT-5 (37.0{\%}), while indicating significant room for future research. All data and code will be publicly released."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="cao-etal-2026-dramasearch">
<titleInfo>
<title>DramaSearch: Benchmarking Multimodal Searching Agents in Ultra-long TV Dramas</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yunning</namePart>
<namePart type="family">Cao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiannan</namePart>
<namePart type="family">Ge</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chuanbin</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lingxi</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinyue</namePart>
<namePart type="family">Huo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pengfei</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kaiwen</namePart>
<namePart type="family">Duan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongtao</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yongdong</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qi</namePart>
<namePart type="family">Tian</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Searching agents have achieved significant progress in NLP, yet their multimodal counterparts are constrained by a lack of suitable evaluation benchmarks requiring reasoning over continuous video, audio, and images. We propose long-form TV dramas as a natural solution and introduce DramaSearch, encompassing 3,600 riddles across 9 dramas totaling 239 hours. Our dataset is generated via a graph-constrained pipeline that processes enriched captions into structured events and multi-granularity evidence views. By employing character-to-view and entity-to-view indexing, we ensure controllable clue disclosure and strictly unique answers. The riddles are categorized into multi-clue, multi-hop, and compositional types, requiring agents to aggregate isolated clues and navigate complex dependency chains. Experiments demonstrate that a Qwen3-8B agent, trained on a held-out set using SFT and RL, achieves an accuracy of 48.3%. This result establishes a strong baseline that performs favorably against zero-shot GPT-5 (37.0%), while indicating significant room for future research. All data and code will be publicly released.</abstract>
<identifier type="citekey">cao-etal-2026-dramasearch</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1313/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>27798</start>
<end>27826</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T DramaSearch: Benchmarking Multimodal Searching Agents in Ultra-long TV Dramas
%A Cao, Yunning
%A Ge, Jiannan
%A Liu, Chuanbin
%A Xie, Lingxi
%A Huo, Xinyue
%A Chen, Pengfei
%A Duan, Kaiwen
%A Xie, Hongtao
%A Zhang, Yongdong
%A Tian, Qi
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F cao-etal-2026-dramasearch
%X Searching agents have achieved significant progress in NLP, yet their multimodal counterparts are constrained by a lack of suitable evaluation benchmarks requiring reasoning over continuous video, audio, and images. We propose long-form TV dramas as a natural solution and introduce DramaSearch, encompassing 3,600 riddles across 9 dramas totaling 239 hours. Our dataset is generated via a graph-constrained pipeline that processes enriched captions into structured events and multi-granularity evidence views. By employing character-to-view and entity-to-view indexing, we ensure controllable clue disclosure and strictly unique answers. The riddles are categorized into multi-clue, multi-hop, and compositional types, requiring agents to aggregate isolated clues and navigate complex dependency chains. Experiments demonstrate that a Qwen3-8B agent, trained on a held-out set using SFT and RL, achieves an accuracy of 48.3%. This result establishes a strong baseline that performs favorably against zero-shot GPT-5 (37.0%), while indicating significant room for future research. All data and code will be publicly released.
%U https://aclanthology.org/2026.findings-emnlp.1313/
%P 27798-27826
Markdown (Informal)
[DramaSearch: Benchmarking Multimodal Searching Agents in Ultra-long TV Dramas](https://aclanthology.org/2026.findings-emnlp.1313/) (Cao et al., Findings 2026)
ACL
- Yunning Cao, Jiannan Ge, Chuanbin Liu, Lingxi Xie, Xinyue Huo, Pengfei Chen, Kaiwen Duan, Hongtao Xie, Yongdong Zhang, and Qi Tian. 2026. DramaSearch: Benchmarking Multimodal Searching Agents in Ultra-long TV Dramas. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 27798–27826, Budapest, Hungary. Association for Computational Linguistics.