Neural Diarization with Non-autoregressive Intermediate Attractors

TL;DR

This paper introduces a novel non-autoregressive neural diarization model that uses intermediate attractors to improve speaker label dependency modeling, leading to enhanced performance and training efficiency.

Contribution

It proposes a new non-autoregressive EEND model with intermediate attractors that refine speaker labels, addressing the lack of label dependency in previous models.

Findings

Improved diarization performance on CALLHOME dataset.

Deeper networks benefit more from intermediate labels.

Enhanced training throughput compared to EEND-EDA.

Abstract

End-to-end neural diarization (EEND) with encoder-decoder-based attractors (EDA) is a promising method to handle the whole speaker diarization problem simultaneously with a single neural network. While the EEND model can produce all frame-level speaker labels simultaneously, it disregards output label dependency. In this work, we propose a novel EEND model that introduces the label dependency between frames. The proposed method generates non-autoregressive intermediate attractors to produce speaker labels at the lower layers and conditions the subsequent layers with these labels. While the proposed model works in a non-autoregressive manner, the speaker labels are refined by referring to the whole sequence of intermediate labels. The experiments with the two-speaker CALLHOME dataset show that the intermediate labels with the proposed non-autoregressive intermediate attractors boost the diarization performance. The proposed method with the deeper network benefits more from the intermediate labels, resulting in better performance and training throughput than EEND-EDA.