Geolocation-Aware Robust Spoken Language Identification

TL;DR

This paper introduces geolocation-aware spoken language identification, enhancing robustness to dialects and accents by integrating geolocation data into SSL models, leading to state-of-the-art accuracy and improved domain generalization.

Contribution

The paper proposes a novel geolocation-aware approach for LID that incorporates geolocation prediction as an auxiliary task and uses it to condition the model's representations.

Findings

Achieves 97.7% accuracy on FLEURS dataset.

Provides 9.7% relative improvement on ML-SUPERB dialect set.

Enhances robustness to intra-language variations and unseen domains.

Abstract

While Self-supervised Learning (SSL) has significantly improved Spoken Language Identification (LID), existing models often struggle to consistently classify dialects and accents of the same language as a unified class. To address this challenge, we propose geolocation-aware LID, a novel approach that incorporates language-level geolocation information into the SSL-based LID model. Specifically, we introduce geolocation prediction as an auxiliary task and inject the predicted vectors into intermediate representations as conditioning signals. This explicit conditioning encourages the model to learn more unified representations for dialectal and accented variations. Experiments across six multilingual datasets demonstrate that our approach improves robustness to intra-language variations and unseen domains, achieving new state-of-the-art accuracy on FLEURS (97.7%) and 9.7% relative improvement on ML-SUPERB 2.0 dialect set.