A Joint and Domain-Adaptive Approach to Spoken Language Understanding

TL;DR

This paper introduces a joint, domain-adaptive model for Spoken Language Understanding that formulates SLU as a constrained generation task, effectively combining intent detection and slot filling with domain adaptation capabilities.

Contribution

It proposes a novel joint and domain-adaptive approach to SLU using constrained generation and dynamic vocabulary, bridging two research lines and enhancing domain adaptability.

Findings

Achieves competitive performance on ASMixed and MTOD datasets.

Demonstrates effective domain adaptation of the joint model.

Bridges joint modeling and domain adaptation in SLU.

Abstract

Spoken Language Understanding (SLU) is composed of two subtasks: intent detection (ID) and slot filling (SF). There are two lines of research on SLU. One jointly tackles these two subtasks to improve their prediction accuracy, and the other focuses on the domain-adaptation ability of one of the subtasks. In this paper, we attempt to bridge these two lines of research and propose a joint and domain adaptive approach to SLU. We formulate SLU as a constrained generation task and utilize a dynamic vocabulary based on domain-specific ontology. We conduct experiments on the ASMixed and MTOD datasets and achieve competitive performance with previous state-of-the-art joint models. Besides, results show that our joint model can be effectively adapted to a new domain.