Spoken Language Understanding on Unseen Tasks With In-Context Learning

TL;DR

This paper proposes a novel fine-tuning method for speech-text large language models that improves their zero-shot and few-shot performance on unseen spoken language understanding tasks without needing task-specific data annotations.

Contribution

It introduces a robust task-agnostic fine-tuning approach using randomized class labels, enhancing LLMs' ability to handle unseen SLU tasks.

Findings

Significant performance improvement on unseen SLU tasks.

Effective zero/few-shot learning without task-specific annotations.

Outperforms standard approaches in robustness and generalization.

Abstract

Spoken language understanding (SLU) tasks involve diverse skills that probe the information extraction, classification and/or generation capabilities of models. In this setting, task-specific training data may not always be available. While traditional task-specific SLU models are unable to cater to such requirements, the speech-text large language models (LLMs) offer a promising alternative with emergent abilities. However, out of-the-box, our evaluations indicate that the zero/few-shot performance of prominent open-source speech-text LLMs on SLU tasks are not up to the mark. In this paper, we introduce a novel approach to robust task-agnostic fine-tuning using randomized class labels. With this proposed fine-tuning, we illustrate that the performance of the speech-text LLMs on an unseen task is significantly improved over standard approaches. Critically, the proposed approach avoids the requirement of task-specific data annotations for enabling new tasks in speech-text LLMs.