Sparse Architectures for Text-Independent Speaker Verification Using Deep Neural Networks

TL;DR

This paper explores structured sparsity in deep neural networks for text-independent speaker verification, demonstrating that pruning can enhance verification accuracy and reduce computational demands.

Contribution

It introduces structured sparsity enforcement in DNNs for speaker verification, showing that pruning can improve performance by mitigating overfitting.

Findings

Sparsity enforcement improves verification accuracy.

Pruned models require less computational power.

Sparsity prevents overfitting in deep networks.

Abstract

Network pruning is of great importance due to the elimination of the unimportant weights or features activated due to the network over-parametrization. Advantages of sparsity enforcement include preventing the overfitting and speedup. Considering a large number of parameters in deep architectures, network compression becomes of critical importance due to the required huge amount of computational power. In this work, we impose structured sparsity for speaker verification which is the validation of the query speaker compared to the speaker gallery. We will show that the mere sparsity enforcement can improve the verification results due to the possible initial overfitting in the network.