Enhancing Audio-Language Models through Self-Supervised Post-Training with Text-Audio Pairs

TL;DR

This paper introduces TeminAL, a two-stage training method that enhances audio-language models with temporal understanding, improving their performance on temporal tasks while maintaining their zero-shot capabilities.

Contribution

The paper proposes TeminAL, a novel two-stage training scheme that instills temporal understanding in contrastive audio-language models without sacrificing their existing capabilities.

Findings

Achieved an average 5.28% performance gain in temporal understanding on ESC-50.

Maintained competitive zero-shot retrieval and classification performance.

Proposed ZSTE, a new evaluation strategy for zero-shot contrastive models.

Abstract

Research on multi-modal contrastive learning strategies for audio and text has rapidly gained interest. Contrastively trained Audio-Language Models (ALMs), such as CLAP, which establish a unified representation across audio and language modalities, have enhanced the efficacy in various subsequent tasks by providing good text aligned audio encoders and vice versa. These improvements are evident in areas like zero-shot audio classification and audio retrieval, among others. However, the ability of these models to understand natural language and temporal relations is still a largely unexplored and open field for research. In this paper, we propose to equip the multi-modal ALMs with temporal understanding without loosing their inherent prior capabilities of audio-language tasks with a temporal instillation method TeminAL. We implement a two-stage training scheme TeminAL A $\&$ B, where the model first learns to differentiate between multiple sounds in TeminAL A, followed by a phase that instills a sense of time, thereby enhancing its temporal understanding in TeminAL B. This approach results in an average performance gain of $5.28\%$ in temporal understanding on the ESC-50 dataset, while the model remains competitive in zero-shot retrieval and classification tasks on the AudioCap/Clotho datasets. We also note the lack of proper evaluation techniques for contrastive ALMs and propose a strategy for evaluating ALMs in zero-shot settings. The general-purpose zero-shot model evaluation strategy ZSTE, is used to evaluate various prior models. ZSTE demonstrates a general strategy to evaluate all ZS contrastive models. The model trained with TeminAL successfully outperforms current models on most downstream tasks.