Supervised Hierarchical Clustering using Graph Neural Networks for Speaker Diarization

TL;DR

This paper introduces E2E-SHARC, a supervised hierarchical clustering method using GNNs for speaker diarization, enabling a single-step, end-to-end process that significantly improves clustering accuracy over traditional methods.

Contribution

The paper presents a novel supervised hierarchical clustering algorithm with GNNs for speaker diarization, allowing end-to-end training and improved performance.

Findings

Achieves 53% relative improvement on AMI dataset.

Achieves 44% relative improvement on Voxconverse dataset.

Enables single-step, end-to-end speaker diarization.

Abstract

Conventional methods for speaker diarization involve windowing an audio file into short segments to extract speaker embeddings, followed by an unsupervised clustering of the embeddings. This multi-step approach generates speaker assignments for each segment. In this paper, we propose a novel Supervised HierArchical gRaph Clustering algorithm (SHARC) for speaker diarization where we introduce a hierarchical structure using Graph Neural Network (GNN) to perform supervised clustering. The supervision allows the model to update the representations and directly improve the clustering performance, thus enabling a single-step approach for diarization. In the proposed work, the input segment embeddings are treated as nodes of a graph with the edge weights corresponding to the similarity scores between the nodes. We also propose an approach to jointly update the embedding extractor and the GNN model to perform end-to-end speaker diarization (E2E-SHARC). During inference, the hierarchical clustering is performed using node densities and edge existence probabilities to merge the segments until convergence. In the diarization experiments, we illustrate that the proposed E2E-SHARC approach achieves 53% and 44% relative improvements over the baseline systems on benchmark datasets like AMI and Voxconverse, respectively.