Saved in:
Bibliographic Details
Main Authors: Fang, Zhihua, He, Liang, Jiang, Weiwu
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.01905
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918479638036480
author Fang, Zhihua
He, Liang
Jiang, Weiwu
author_facet Fang, Zhihua
He, Liang
Jiang, Weiwu
contents For the speaker-controlled spoken language identification task proposed in the TidyLang Challenge 2026, this paper proposes a language identification method based on pre-trained models and margin-based losses. The proposed method adopts a pre-trained ECAPA-TDNN as the feature encoder and incorporates margin-based losses to enhance the discriminative ability of language representations, thereby improving inter-class separability and reducing the interference of non-linguistic factors such as speaker characteristics. Experimental results on the Tidy-X dataset show that the proposed method achieves 85.95% macro accuracy and 90.96% micro accuracy on the language identification task and 17.08% equal error rate (EER) on the verification task. Compared with the official baseline, the macro accuracy improves by 45.7%, the micro accuracy improves by 15.2%, and the EER is reduced by approximately 50.8%, demonstrating the effectiveness of the proposed method. The code will be released at https://github.com/PunkMale/TidyLang2026.
format Preprint
id arxiv_https___arxiv_org_abs_2605_01905
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Spoken Language Identification with Pre-trained Models and Margin Loss
Fang, Zhihua
He, Liang
Jiang, Weiwu
Sound
Computation and Language
For the speaker-controlled spoken language identification task proposed in the TidyLang Challenge 2026, this paper proposes a language identification method based on pre-trained models and margin-based losses. The proposed method adopts a pre-trained ECAPA-TDNN as the feature encoder and incorporates margin-based losses to enhance the discriminative ability of language representations, thereby improving inter-class separability and reducing the interference of non-linguistic factors such as speaker characteristics. Experimental results on the Tidy-X dataset show that the proposed method achieves 85.95% macro accuracy and 90.96% micro accuracy on the language identification task and 17.08% equal error rate (EER) on the verification task. Compared with the official baseline, the macro accuracy improves by 45.7%, the micro accuracy improves by 15.2%, and the EER is reduced by approximately 50.8%, demonstrating the effectiveness of the proposed method. The code will be released at https://github.com/PunkMale/TidyLang2026.
title Spoken Language Identification with Pre-trained Models and Margin Loss
topic Sound
Computation and Language
url https://arxiv.org/abs/2605.01905