Towards Speaker Age Estimation with Label Distribution Learning

TL;DR

This paper introduces a label distribution learning approach for speaker age estimation that captures label ambiguity, combining classification and regression to improve accuracy and robustness, demonstrated by significant MAE reduction on real-world data.

Contribution

The paper proposes a novel LDL-based method for speaker age estimation that models age labels as distributions, effectively handling label ambiguity and improving performance.

Findings

Outperforms baseline methods with a 10% MAE reduction on real-world data.

Combines classification and regression for more robust age estimation.

Demonstrates effectiveness on public and real-world datasets.

Abstract

Existing methods for speaker age estimation usually treat it as a multi-class classification or a regression problem. However, precise age identification remains a challenge due to label ambiguity, \emph{i.e.}, utterances from adjacent age of the same person are often indistinguishable. To address this, we utilize the ambiguous information among the age labels, convert each age label into a discrete label distribution and leverage the label distribution learning (LDL) method to fit the data. For each audio data sample, our method produces a age distribution of its speaker, and on top of the distribution we also perform two other tasks: age prediction and age uncertainty minimization. Therefore, our method naturally combines the age classification and regression approaches, which enhances the robustness of our method. We conduct experiments on the public NIST SRE08-10 dataset and a real-world dataset, which exhibit that our method outperforms baseline methods by a relatively large margin, yielding a 10\% reduction in terms of mean absolute error (MAE) on a real-world dataset.