@inproceedings{chaichana-etal-2026-extending,
title = "Extending Audio Context for Long-Form Understanding in Large Audio-Language Models",
author = "Chaichana, Yuatyong and
Taveekitworachai, Pittawat and
Sirichotedumrong, Warit and
Manakul, Potsawee and
Pipatanakul, Kunat",
editor = "Demberg, Vera and
Inui, Kentaro and
Marquez, Llu{\'i}s",
booktitle = "Proceedings of the 19th Conference of the {E}uropean Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = mar,
year = "2026",
address = "Rabat, Morocco",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.eacl-long.286/",
pages = "6046--6066",
ISBN = "979-8-89176-380-7",
abstract = "Large Audio-Language Models (LALMs) are often constrained by short audio context windows, even when their text backbones support long contexts, limiting long-form audio understanding. Prior work has introduced context-extension methods (e.g. YaRN) on unimodal LLMs, yet their application to LALMs remains unexplored. First, building on RoPE-based context extension, we introduce \textbf{Partial YaRN}, a training-free, modality-decoupled extension method that modifies only audio token positions, leaving text positions intact to preserve the base LLM{'}s text capabilities. Second, we propose Virtual Longform Audio Training (\textbf{VLAT}), a training strategy that extends Partial YaRN into a training-time positional augmentation. VLAT simulates diverse audio lengths during training, enabling generalization to inputs far longer than those seen in training. Our experiments on SALMONN and Qwen2-Audio confirm that Partial YaRN outperforms the original models across wide range of settings, and VLAT provides substantial performance improvement on long audio of unseen lengths."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chaichana-etal-2026-extending">
<titleInfo>
<title>Extending Audio Context for Long-Form Understanding in Large Audio-Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yuatyong</namePart>
<namePart type="family">Chaichana</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pittawat</namePart>
<namePart type="family">Taveekitworachai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Warit</namePart>
<namePart type="family">Sirichotedumrong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Potsawee</namePart>
<namePart type="family">Manakul</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kunat</namePart>
<namePart type="family">Pipatanakul</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-03</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Vera</namePart>
<namePart type="family">Demberg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kentaro</namePart>
<namePart type="family">Inui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lluís</namePart>
<namePart type="family">Marquez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Rabat, Morocco</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-380-7</identifier>
</relatedItem>
<abstract>Large Audio-Language Models (LALMs) are often constrained by short audio context windows, even when their text backbones support long contexts, limiting long-form audio understanding. Prior work has introduced context-extension methods (e.g. YaRN) on unimodal LLMs, yet their application to LALMs remains unexplored. First, building on RoPE-based context extension, we introduce Partial YaRN, a training-free, modality-decoupled extension method that modifies only audio token positions, leaving text positions intact to preserve the base LLM’s text capabilities. Second, we propose Virtual Longform Audio Training (VLAT), a training strategy that extends Partial YaRN into a training-time positional augmentation. VLAT simulates diverse audio lengths during training, enabling generalization to inputs far longer than those seen in training. Our experiments on SALMONN and Qwen2-Audio confirm that Partial YaRN outperforms the original models across wide range of settings, and VLAT provides substantial performance improvement on long audio of unseen lengths.</abstract>
<identifier type="citekey">chaichana-etal-2026-extending</identifier>
<location>
<url>https://aclanthology.org/2026.eacl-long.286/</url>
</location>
<part>
<date>2026-03</date>
<extent unit="page">
<start>6046</start>
<end>6066</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Extending Audio Context for Long-Form Understanding in Large Audio-Language Models
%A Chaichana, Yuatyong
%A Taveekitworachai, Pittawat
%A Sirichotedumrong, Warit
%A Manakul, Potsawee
%A Pipatanakul, Kunat
%Y Demberg, Vera
%Y Inui, Kentaro
%Y Marquez, Lluís
%S Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 March
%I Association for Computational Linguistics
%C Rabat, Morocco
%@ 979-8-89176-380-7
%F chaichana-etal-2026-extending
%X Large Audio-Language Models (LALMs) are often constrained by short audio context windows, even when their text backbones support long contexts, limiting long-form audio understanding. Prior work has introduced context-extension methods (e.g. YaRN) on unimodal LLMs, yet their application to LALMs remains unexplored. First, building on RoPE-based context extension, we introduce Partial YaRN, a training-free, modality-decoupled extension method that modifies only audio token positions, leaving text positions intact to preserve the base LLM’s text capabilities. Second, we propose Virtual Longform Audio Training (VLAT), a training strategy that extends Partial YaRN into a training-time positional augmentation. VLAT simulates diverse audio lengths during training, enabling generalization to inputs far longer than those seen in training. Our experiments on SALMONN and Qwen2-Audio confirm that Partial YaRN outperforms the original models across wide range of settings, and VLAT provides substantial performance improvement on long audio of unseen lengths.
%U https://aclanthology.org/2026.eacl-long.286/
%P 6046-6066
Markdown (Informal)
[Extending Audio Context for Long-Form Understanding in Large Audio-Language Models](https://aclanthology.org/2026.eacl-long.286/) (Chaichana et al., EACL 2026)
ACL
- Yuatyong Chaichana, Pittawat Taveekitworachai, Warit Sirichotedumrong, Potsawee Manakul, and Kunat Pipatanakul. 2026. Extending Audio Context for Long-Form Understanding in Large Audio-Language Models. In Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6046–6066, Rabat, Morocco. Association for Computational Linguistics.