Adaptation Algorithms for Neural Network-Based Speech Recognition: An Overview

TL;DR

This paper provides a comprehensive overview of adaptation algorithms for neural network-based speech recognition, analyzing various methods like embeddings, parameter adaptation, and data augmentation, and summarizing their performance improvements.

Contribution

It offers a structured classification and meta-analysis of adaptation techniques for neural speech recognition systems, highlighting their relative effectiveness.

Findings

Embeddings and parameter adaptation are effective for speaker and domain adaptation.

Meta-analysis shows significant error rate reductions with certain adaptation methods.

The overview identifies gaps and future directions in speech recognition adaptation research.

Abstract

We present a structured overview of adaptation algorithms for neural network-based speech recognition, considering both hybrid hidden Markov model / neural network systems and end-to-end neural network systems, with a focus on speaker adaptation, domain adaptation, and accent adaptation. The overview characterizes adaptation algorithms as based on embeddings, model parameter adaptation, or data augmentation. We present a meta-analysis of the performance of speech recognition adaptation algorithms, based on relative error rate reductions as reported in the literature.