Interrelate Training and Searching: A Unified Online Clustering Framework for Speaker Diarization

TL;DR

This paper introduces a unified online clustering framework for speaker diarization that integrates training and searching processes, significantly improving real-time clustering accuracy.

Contribution

It proposes a novel interactive framework combining clustering-guided recurrent training and truncated beam searching clustering for online speaker diarization.

Findings

Achieved 14.48% DER on AISHELL-4 with low latency

Outperformed offline agglomerative hierarchical clustering

Introduced cluster-aware training for embedding extractors

Abstract

For online speaker diarization, samples arrive incrementally, and the overall distribution of the samples is invisible. Moreover, in most existing clustering-based methods, the training objective of the embedding extractor is not designed specially for clustering. To improve online speaker diarization performance, we propose a unified online clustering framework, which provides an interactive manner between embedding extractors and clustering algorithms. Specifically, the framework consists of two highly coupled parts: clustering-guided recurrent training (CGRT) and truncated beam searching clustering (TBSC). The CGRT introduces the clustering algorithm into the training process of embedding extractors, which could provide not only cluster-aware information for the embedding extractor, but also crucial parameters for the clustering process afterward. And with these parameters, which contain preliminary information of the metric space, the TBSC penalizes the probability score of each cluster, in order to output more accurate clustering results in online fashion with low latency. With the above innovations, our proposed online clustering system achieves 14.48\% DER with collar 0.25 at 2.5s latency on the AISHELL-4, while the DER of the offline agglomerative hierarchical clustering is 14.57\%.