@inproceedings{kahatapitiya-etal-2025-language,
title = "Language Repository for Long Video Understanding",
author = "Kahatapitiya, Kumara and
Ranasinghe, Kanchana and
Park, Jongwoo and
Ryoo, Michael S",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.294/",
doi = "10.18653/v1/2025.findings-acl.294",
pages = "5627--5646",
ISBN = "979-8-89176-256-5",
abstract = "Language has become a prominent modality in computer vision with the rise of LLMs. Despite supporting long context-lengths, their effectiveness in handling long-term information gradually declines with input length. This becomes critical, especially in applications such as long-form video understanding. In this paper, we introduce a Language Repository (LangRepo) for LLMs, that maintains concise and structured information as an interpretable (i.e., all-textual) representation. Our repository is updated iteratively based on multi-scale video chunks. We introduce write and read operations that focus on pruning redundancies in text, and extracting information at various temporal scales. The proposed framework is evaluated on zero-shot visual question-answering benchmarks, showing state-of-the-art performance at its scale. Our code is available at https://github.com/kkahatapitiya/LangRepo."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kahatapitiya-etal-2025-language">
<titleInfo>
<title>Language Repository for Long Video Understanding</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kumara</namePart>
<namePart type="family">Kahatapitiya</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kanchana</namePart>
<namePart type="family">Ranasinghe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jongwoo</namePart>
<namePart type="family">Park</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Michael</namePart>
<namePart type="given">S</namePart>
<namePart type="family">Ryoo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>Language has become a prominent modality in computer vision with the rise of LLMs. Despite supporting long context-lengths, their effectiveness in handling long-term information gradually declines with input length. This becomes critical, especially in applications such as long-form video understanding. In this paper, we introduce a Language Repository (LangRepo) for LLMs, that maintains concise and structured information as an interpretable (i.e., all-textual) representation. Our repository is updated iteratively based on multi-scale video chunks. We introduce write and read operations that focus on pruning redundancies in text, and extracting information at various temporal scales. The proposed framework is evaluated on zero-shot visual question-answering benchmarks, showing state-of-the-art performance at its scale. Our code is available at https://github.com/kkahatapitiya/LangRepo.</abstract>
<identifier type="citekey">kahatapitiya-etal-2025-language</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.294</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.294/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>5627</start>
<end>5646</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Language Repository for Long Video Understanding
%A Kahatapitiya, Kumara
%A Ranasinghe, Kanchana
%A Park, Jongwoo
%A Ryoo, Michael S.
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F kahatapitiya-etal-2025-language
%X Language has become a prominent modality in computer vision with the rise of LLMs. Despite supporting long context-lengths, their effectiveness in handling long-term information gradually declines with input length. This becomes critical, especially in applications such as long-form video understanding. In this paper, we introduce a Language Repository (LangRepo) for LLMs, that maintains concise and structured information as an interpretable (i.e., all-textual) representation. Our repository is updated iteratively based on multi-scale video chunks. We introduce write and read operations that focus on pruning redundancies in text, and extracting information at various temporal scales. The proposed framework is evaluated on zero-shot visual question-answering benchmarks, showing state-of-the-art performance at its scale. Our code is available at https://github.com/kkahatapitiya/LangRepo.
%R 10.18653/v1/2025.findings-acl.294
%U https://aclanthology.org/2025.findings-acl.294/
%U https://doi.org/10.18653/v1/2025.findings-acl.294
%P 5627-5646
Markdown (Informal)
[Language Repository for Long Video Understanding](https://aclanthology.org/2025.findings-acl.294/) (Kahatapitiya et al., Findings 2025)
ACL
- Kumara Kahatapitiya, Kanchana Ranasinghe, Jongwoo Park, and Michael S Ryoo. 2025. Language Repository for Long Video Understanding. In Findings of the Association for Computational Linguistics: ACL 2025, pages 5627–5646, Vienna, Austria. Association for Computational Linguistics.