Novel Architectures for Unsupervised Information Bottleneck based Speaker Diarization of Meetings

TL;DR

This paper introduces novel unsupervised speaker diarization architectures that improve segmentation initialization and incorporate speaker discriminative features, resulting in significant performance gains on standard meeting datasets.

Contribution

It proposes a varying length segment initialization technique and a Two-Pass Information Bottleneck framework that together enhance unsupervised speaker diarization performance.

Findings

Achieved 3.9% and 4.7% absolute improvement on NIST and AMI datasets.

Demonstrated the effectiveness of combining segment initialization with discriminative features.

Improved baseline IB-based diarization with novel initialization and two-pass approach.

Abstract

Speaker diarization is an important problem that is topical, and is especially useful as a preprocessor for conversational speech related applications. The objective of this paper is two-fold: (i) segment initialization by uniformly distributing speaker information across the initial segments, and (ii) incorporating speaker discriminative features within the unsupervised diarization framework. In the first part of the work, a varying length segment initialization technique for Information Bottleneck (IB) based speaker diarization system using phoneme rate as the side information is proposed. This initialization distributes speaker information uniformly across the segments and provides a better starting point for IB based clustering. In the second part of the work, we present a Two-Pass Information Bottleneck (TPIB) based speaker diarization system that incorporates speaker discriminative features during the process of diarization. The TPIB based speaker diarization system has shown improvement over the baseline IB based system. During the first pass of the TPIB system, a coarse segmentation is performed using IB based clustering. The alignments obtained are used to generate speaker discriminative features using a shallow feed-forward neural network and linear discriminant analysis. The discriminative features obtained are used in the second pass to obtain the final speaker boundaries. In the final part of the paper, variable segment initialization is combined with the TPIB framework. This leverages the advantages of better segment initialization and speaker discriminative features that results in an additional improvement in performance. An evaluation on standard meeting datasets shows that a significant absolute improvement of 3.9% and 4.7% is obtained on the NIST and AMI datasets, respectively.