Improving Neural Diarization through Speaker Attribute Attractors and Local Dependency Modeling

TL;DR

This paper advances neural speaker diarization by incorporating detailed speaker attribute representations and local dependency modeling using conformers, resulting in improved performance on benchmark datasets.

Contribution

It introduces a multi-stage speaker attribute representation and replaces transformers with conformers to enhance local dependency modeling in diarization.

Findings

Improved diarization accuracy on CALLHOME dataset.

Enhanced modeling of speaker attributes through multi-stage representations.

Effective integration of conformers for local dependency capture.

Abstract

In recent years, end-to-end approaches have made notable progress in addressing the challenge of speaker diarization, which involves segmenting and identifying speakers in multi-talker recordings. One such approach, Encoder-Decoder Attractors (EDA), has been proposed to handle variable speaker counts as well as better guide the network during training. In this study, we extend the attractor paradigm by moving beyond direct speaker modeling and instead focus on representing more detailed `speaker attributes' through a multi-stage process of intermediate representations. Additionally, we enhance the architecture by replacing transformers with conformers, a convolution-augmented transformer, to model local dependencies. Experiments demonstrate improved diarization performance on the CALLHOME dataset.