Attention-based conditioning methods using variable frame rate for style-robust speaker verification

TL;DR

This paper introduces an entropy-based variable frame rate conditioning method for self-attention in speaker verification, improving robustness to speaking style variations across multiple datasets.

Contribution

It proposes a novel entropy-based conditioning vector for self-attention, enhancing speaker embedding robustness to style variations in text-independent verification.

Findings

Significant improvements over baseline in 12/23 tasks

Outperforms unconditioned self-attention in 9/23 tasks

Effective in multi-speaker scenarios like SITW

Abstract

We propose an approach to extract speaker embeddings that are robust to speaking style variations in text-independent speaker verification. Typically, speaker embedding extraction includes training a DNN for speaker classification and using the bottleneck features as speaker representations. Such a network has a pooling layer to transform frame-level to utterance-level features by calculating statistics over all utterance frames, with equal weighting. However, self-attentive embeddings perform weighted pooling such that the weights correspond to the importance of the frames in a speaker classification task. Entropy can capture acoustic variability due to speaking style variations. Hence, an entropy-based variable frame rate vector is proposed as an external conditioning vector for the self-attention layer to provide the network with information that can address style effects. This work explores five different approaches to conditioning. The best conditioning approach, concatenation with gating, provided statistically significant improvements over the x-vector baseline in 12/23 tasks and was the same as the baseline in 11/23 tasks when using the UCLA speaker variability database. It also significantly outperformed self-attention without conditioning in 9/23 tasks and was worse in 1/23. The method also showed significant improvements in multi-speaker scenarios of SITW.