Cross-lingual Embedding Clustering for Hierarchical Softmax in Low-Resource Multilingual Speech Recognition

TL;DR

This paper introduces a cross-lingual embedding clustering method to improve hierarchical Softmax decoding in low-resource multilingual speech recognition, leading to better accuracy across multiple languages.

Contribution

It proposes a novel cross-lingual embedding clustering approach for hierarchical Softmax, addressing limitations of previous methods and enhancing multilingual ASR performance.

Findings

Improved accuracy in low-resource multilingual ASR

Effective sharing of token representations across languages

Outperforms Huffman-based H-Softmax in experiments

Abstract

We present a novel approach centered on the decoding stage of Automatic Speech Recognition (ASR) that enhances multilingual performance, especially for low-resource languages. It utilizes a cross-lingual embedding clustering method to construct a hierarchical Softmax (H-Softmax) decoder, which enables similar tokens across different languages to share similar decoder representations. It addresses the limitations of the previous Huffman-based H-Softmax method, which relied on shallow features in token similarity assessments. Through experiments on a downsampled dataset of 15 languages, we demonstrate the effectiveness of our approach in improving low-resource multilingual ASR accuracy.