Synthio: Augmenting Small-Scale Audio Classification Datasets with Synthetic Data

TL;DR

Synthio introduces a novel synthetic data augmentation method for small-scale audio classification datasets, leveraging text-to-audio diffusion models aligned with dataset characteristics and enhanced by large language model-generated captions, significantly improving accuracy.

Contribution

The paper presents a new approach combining preference optimization and LLM-generated captions to produce diverse, acoustically consistent synthetic audio data for small datasets.

Findings

Outperforms baselines by up to 39% in accuracy.

Effective in ten datasets with limited data scenarios.

Uses weakly-captioned AudioSet for training T2A models.

Abstract

We present Synthio, a novel approach for augmenting small-scale audio classification datasets with synthetic data. Our goal is to improve audio classification accuracy with limited labeled data. Traditional data augmentation techniques, which apply artificial transformations (e.g., adding random noise or masking segments), struggle to create data that captures the true diversity present in real-world audios. To address this shortcoming, we propose to augment the dataset with synthetic audio generated from text-to-audio (T2A) diffusion models. However, synthesizing effective augmentations is challenging because not only should the generated data be acoustically consistent with the underlying small-scale dataset, but they should also have sufficient compositional diversity. To overcome the first challenge, we align the generations of the T2A model with the small-scale dataset using preference optimization. This ensures that the acoustic characteristics of the generated data remain consistent with the small-scale dataset. To address the second challenge, we propose a novel caption generation technique that leverages the reasoning capabilities of Large Language Models to (1) generate diverse and meaningful audio captions and (2) iteratively refine their quality. The generated captions are then used to prompt the aligned T2A model. We extensively evaluate Synthio on ten datasets and four simulated limited-data settings. Results indicate our method consistently outperforms all baselines by 0.1%-39% using a T2A model trained only on weakly-captioned AudioSet.