Enhancing Temporal Understanding in Audio Question Answering for Large Audio Language Models

TL;DR

This paper improves temporal reasoning in Large Audio Language Models for audio question answering by introducing data augmentation, curriculum fine-tuning, and on-device deployment, enhancing their practical applicability.

Contribution

It proposes a novel data augmentation method and curriculum fine-tuning strategy to enhance temporal reasoning in LALMs for audio QA tasks.

Findings

Enhanced temporal reasoning performance on benchmark datasets

Successful on-device CPU inference demonstration

Improved model specialization without performance trade-offs

Abstract

The Audio Question Answering (AQA) task includes audio event classification, audio captioning, and open-ended reasoning. Recently, AQA has garnered attention due to the advent of Large Audio Language Models (LALMs). Current literature focuses on constructing LALMs by integrating audio encoders with text-only Large Language Models (LLMs) through a projection module. While LALMs excel in general audio understanding, they are limited in temporal reasoning, which may hinder their commercial applications and on-device deployment. This paper addresses these challenges and limitations in audio temporal reasoning. First, we introduce a data augmentation technique for generating reliable audio temporal questions and answers using an LLM. Second, we perform a further fine-tuning of an existing baseline using curriculum learning strategy to specialize in temporal reasoning without compromising performance on fine-tuned tasks. We demonstrate the performance of our model using state-of-the-art LALMs on public audio benchmark datasets. Third, we implement our AQA model on-device locally and investigate its CPU inference for edge applications.