Data-Balanced Curriculum Learning for Audio Question Answering

TL;DR

This paper introduces a data-balanced curriculum learning approach for audio question answering that improves model accuracy by addressing dataset imbalance and training stability issues.

Contribution

It combines curriculum learning with statistical data balancing and guided decoding, a novel approach for enhancing AQA performance.

Findings

Achieves 64.2% accuracy on DCASE 2025 benchmark.

Improves accuracy by 11.7% over baseline models.

Effectively handles dataset imbalance and training instability.

Abstract

Audio question answering (AQA) requires models to understand acoustic content and perform complex reasoning. Current models struggle with dataset imbalances and unstable training dynamics. This work combines curriculum learning with statistical data balancing to address these challenges. The method labels question difficulty using language models, then trains progressively from easy to hard examples. Statistical filtering removes overrepresented audio categories, and guided decoding constrains outputs to valid multiple-choice formats. Experiments on the DCASE 2025 training set and five additional public datasets show that data curation improves accuracy by 11.7% over baseline models, achieving 64.2% on the DCASE 2025 benchmark.