Skip to content

Companion repository to the paper "On the calibration of powerset speaker diarization models" published at Interspeech 2024

Notifications You must be signed in to change notification settings

FrenchKrab/IS2024-powerset-calibration

Repository files navigation

On the calibration of powerset speaker diarization models

Alexis Plaquet and Hervé Bredin
Proc. InterSpeech 2024.

End-to-end neural diarization models have usually relied on a multilabel-classification formulation of the speaker diarization problem. Recently, we proposed a powerset multiclass formulation that has beaten the state-of-the-art on multiple datasets. In this paper, we propose to study the calibration of a powerset speaker diarization model, and explore some of its uses. We study the calibration in-domain, as well as out-of-domain, and explore the data in low-confidence regions. The reliability of model confidence is then tested in practice: we use the confidence of the pretrained model to selectively create training and validation subsets out of unannotated data, and compare this to random selection. We find that top-label confidence can be used to reliably predict high-error regions. Moreover, training on low-confidence regions provides a better calibrated model, and validating on low-confidence regions can be more annotation-efficient than random regions.

📄 Read the paper (ISCA)

🌐 Browse the companion website

🐍 Install the python package

Citations

To be added.

About

Companion repository to the paper "On the calibration of powerset speaker diarization models" published at Interspeech 2024

Topics

Resources

Stars

Watchers

Forks

Languages