CA-MHFA: A Context-Aware Multi-Head Factorized Attentive Pooling for SSL-Based Speaker Verification

TL;DR

This paper introduces CA-MHFA, a lightweight, context-aware pooling method for SSL-based speaker verification that improves accuracy, efficiency, and generalization across tasks by modeling local temporal dependencies effectively.

Contribution

The paper proposes CA-MHFA, a novel context-aware multi-head attentive pooling framework that enhances SSL-based speaker verification by capturing local dependencies with fewer parameters.

Findings

Achieves state-of-the-art EERs on VoxCeleb benchmarks

Outperforms complex models like WavLM-TDNN with fewer parameters

Demonstrates strong generalization across multiple tasks

Abstract

Self-supervised learning (SSL) models for speaker verification (SV) have gained significant attention in recent years. However, existing SSL-based SV systems often struggle to capture local temporal dependencies and generalize across different tasks. In this paper, we propose context-aware multi-head factorized attentive pooling (CA-MHFA), a lightweight framework that incorporates contextual information from surrounding frames. CA-MHFA leverages grouped, learnable queries to effectively model contextual dependencies while maintaining efficiency by sharing keys and values across groups. Experimental results on the VoxCeleb dataset show that CA-MHFA achieves EERs of 0.42\%, 0.48\%, and 0.96\% on Vox1-O, Vox1-E, and Vox1-H, respectively, outperforming complex models like WavLM-TDNN with fewer parameters and faster convergence. Additionally, CA-MHFA demonstrates strong generalization across multiple SSL models and tasks, including emotion recognition and anti-spoofing, highlighting its robustness and versatility.