LaDA: Latent Dialogue Action For Zero-shot Cross-lingual Neural Network Language Modeling

TL;DR

This paper introduces LaDA, a latent dialogue action layer that enhances zero-shot and few-shot cross-lingual spoken language understanding by optimizing decoding strategies, especially for distant languages with different scripts and structures.

Contribution

LaDA is the first to use latent variables for optimizing cross-lingual SLU decoding, significantly improving intent detection and slot filling in distant languages.

Findings

Achieves state-of-the-art results on public datasets

Improves handling of complex multilingual intent and slot values

Effective in zero-shot and few-shot adaptation scenarios

Abstract

Cross-lingual adaptation has proven effective in spoken language understanding (SLU) systems with limited resources. Existing methods are frequently unsatisfactory for intent detection and slot filling, particularly for distant languages that differ significantly from the source language in scripts, morphology, and syntax. Latent Dialogue Action (LaDA) layer is proposed to optimize decoding strategy in order to address the aforementioned issues. The model consists of an additional layer of latent dialogue action. It enables our model to improve a system's capability of handling conversations with complex multilingual intent and slot values of distant languages. To the best of our knowledge, this is the first exhaustive investigation of the use of latent variables for optimizing cross-lingual SLU policy during the decode stage. LaDA obtains state-of-the-art results on public datasets for both zero-shot and few-shot adaptation.