DiaCorrect: Error Correction Back-end For Speaker Diarization

TL;DR

DiaCorrect is a novel error correction framework that refines speaker diarization outputs by leveraging a dual-encoder and transformer-based decoder, significantly reducing errors in telephony data.

Contribution

It introduces a new error correction approach for speaker diarization inspired by speech recognition techniques, utilizing a dual-encoder and transformer decoder architecture.

Findings

Effective improvement on 2-speaker telephony data

Reduces diarization errors significantly

Open-source implementation available

Abstract

In this work, we propose an error correction framework, named DiaCorrect, to refine the output of a diarization system in a simple yet effective way. This method is inspired by error correction techniques in automatic speech recognition. Our model consists of two parallel convolutional encoders and a transform-based decoder. By exploiting the interactions between the input recording and the initial system's outputs, DiaCorrect can automatically correct the initial speaker activities to minimize the diarization errors. Experiments on 2-speaker telephony data show that the proposed DiaCorrect can effectively improve the initial model's results. Our source code is publicly available at https://github.com/BUTSpeechFIT/diacorrect.