Acoustic Prompt Tuning: Empowering Large Language Models with Audition Capabilities

TL;DR

This paper introduces Acoustic Prompt Tuning (APT), a method to extend large language and vision models to the audio domain using soft prompts and curriculum learning, enabling diverse audio tasks without fine-tuning.

Contribution

The work proposes a novel adapter, APT, that injects audio embeddings into LLMs and VLMs, along with a curriculum learning strategy and interleaved audio-text inputs for versatile audio understanding.

Findings

APT-LLMs achieve competitive results on various audio tasks.

The method extends VLMs to audio without fine-tuning.

Introduces natural language audio reasoning (NLAR) task.

Abstract

The auditory system plays a substantial role in shaping the overall human perceptual experience. While prevailing large language models (LLMs) and visual language models (VLMs) have shown their promise in solving a wide variety of language and vision understanding tasks, only a few of them can be generalised to the audio domain without compromising their domain-specific capability. In this work, we introduce Acoustic Prompt Tuning (APT), a new adapter extending LLMs and VLMs to the audio domain by injecting audio embeddings to the input of LLMs, namely soft prompting. Specifically, APT applies an instruction-aware audio aligner to generate soft prompts, conditioned on both input text and sounds, as the inputs to the language model. To mitigate data scarcity in the audio domain, a curriculum learning strategy is proposed by formulating diverse audio tasks in a sequential manner. Moreover, we improve the audio language model by using interleaved audio-text embeddings as the input sequence. In this improved model, zero constraints are imposed on the input format, thus it is capable of tackling diverse modelling tasks, such as few-shot audio classification and audio comparison. To further evaluate the advanced ability of the audio networks, we introduce natural language audio reasoning (NLAR), a new task that analyses two audio clips by comparison and summarisation. Experiments show that APT-enhanced LLMs (namely APT-LLMs) achieve competitive results compared to the expert models (i.e., the networks trained on the target datasets) across various tasks. We finally demonstrate APT's ability in extending frozen VLMs to the audio domain without fine-tuning, achieving promising results in audio-visual question and answering. Our code and model weights will be released at https://github.com/JinhuaLiang/APT