Extending Audio Context for Long-Form Understanding in Large Audio-Language Models

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Chaichana, Yuatyong, Taveekitworachai, Pittawat, Sirichotedumrong, Warit, Manakul, Potsawee, Pipatanakul, Kunat
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912836263870464
author Chaichana, Yuatyong
Taveekitworachai, Pittawat
Sirichotedumrong, Warit
Manakul, Potsawee
Pipatanakul, Kunat
author_facet Chaichana, Yuatyong
Taveekitworachai, Pittawat
Sirichotedumrong, Warit
Manakul, Potsawee
Pipatanakul, Kunat
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2510_15231
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Extending Audio Context for Long-Form Understanding in Large Audio-Language Models
Chaichana, Yuatyong
Taveekitworachai, Pittawat
Sirichotedumrong, Warit
Manakul, Potsawee
Pipatanakul, Kunat
Computation and Language
Artificial Intelligence
Sound
Audio and Speech Processing
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.
title Extending Audio Context for Long-Form Understanding in Large Audio-Language Models
topic Computation and Language
Artificial Intelligence
Sound
Audio and Speech Processing
url https://arxiv.org/abs/2510.15231