A Study on Decoupled Probabilistic Linear Discriminant Analysis

TL;DR

This paper introduces a decoupled approach to probabilistic linear discriminant analysis (PLDA) that combines global simplicity with local complexity, improving speaker verification performance on raw vectors but with limitations after normalization.

Contribution

It proposes a novel decoupling method for PLDA that separates global and local modeling, enhancing flexibility while maintaining generalization.

Findings

Decoupled PLDA outperforms vanilla PLDA on raw speaker vectors.

The advantage diminishes after length normalization of vectors.

Future work suggested on non-linear local models.

Abstract

Probabilistic linear discriminant analysis (PLDA) has broad application in open-set verification tasks, such as speaker verification. A key concern for PLDA is that the model is too simple (linear Gaussian) to deal with complicated data; however, the simplicity by itself is a major advantage of PLDA, as it leads to desirable generalization. An interesting research therefore is how to improve modeling capacity of PLDA while retaining the simplicity. This paper presents a decoupling approach, which involves a global model that is simple and generalizable, and a local model that is complex and expressive. While the global model holds a bird view on the entire data, the local model represents the details of individual classes. We conduct a preliminary study towards this direction and investigate a simple decoupling model including both the global and local models. The new model, which we call decoupled PLDA, is tested on a speaker verification task. Experimental results show that it consistently outperforms the vanilla PLDA when the model is based on raw speaker vectors. However, when the speaker vectors are processed by length normalization, the advantage of decoupled PLDA will be largely lost, suggesting future research on non-linear local models.