Gender domain adaptation for automatic speech recognition task

TL;DR

This study investigates gender-specific finetuning of Transformer-based acoustic models for speech recognition, finding modest improvements in word error rate when adapting models to gender and accented speech, with limited gains from voice embedding concatenation.

Contribution

It explores gender-specific finetuning strategies for Transformer acoustic models and evaluates their effectiveness on accented speech datasets.

Findings

Up to 5% lower WER on male subset without freezing layers.

Gender-specific models outperform models trained on full dataset by 1-2%.

Voice embedding concatenation shows limited accuracy improvement.

Abstract

This paper is focused on the finetuning of acoustic models for speaker adaptation goals on a given gender. We pretrained the Transformer baseline model on Librispeech-960 and conduct experiments with finetuning on the gender-specific test subsets and. In general, we do not obtain essential WER reduction by finetuning techniques by this approach. We achieved up to ~5% lower word error rate on the male subset and 3% on the female subset if the layers in the encoder and decoder are not frozen, but the tuning is started from the last checkpoints. Moreover, we adapted our base model on the full L2 Arctic dataset of accented speech and fine-tuned it for particular speakers and male and female genders separately. The models trained on the gender subsets obtained 1-2% higher accuracy when compared to the model tuned on the whole L2 Arctic dataset. Finally, we tested the concatenation of the pretrained x-vector voice embeddings and embeddings from a conventional encoder, but its gain in accuracy is not significant.