Research

Paper

AI LLM March 11, 2026

Speaker Verification with Speech-Aware LLMs: Evaluation and Augmentation

Authors

Thomas Thebaud, Yuzhe Wang, Laureano Moro-Velazquez, Jesus Villalba-Lopez, Najim Dehak

Abstract

Speech-aware large language models (LLMs) can accept speech inputs, yet their training objectives largely emphasize linguistic content or specific fields such as emotions or the speaker's gender, leaving it unclear whether they encode speaker identity. First, we propose a model-agnostic scoring protocol that produces continuous verification scores for both API-only and open-weight models, using confidence scores or log-likelihood ratios from the Yes/No token probabilities. Using this protocol, we benchmark recent speech-aware LLMs and observe weak speaker discrimination (EERs above 20% on VoxCeleb1). Second, we introduce a lightweight augmentation that equips an LLM with ASV capability by injecting frozen ECAPA-TDNN speaker embeddings through a learned projection and training only LoRA adapters. On TinyLLaMA-1.1B, the resulting ECAPA-LLM achieves 1.03% EER on VoxCeleb1-E, approaching a dedicated speaker verification system while preserving a natural-language interface.

Metadata

arXiv ID: 2603.10827
Provider: ARXIV
Primary Category: cs.SD
Published: 2026-03-11
Fetched: 2026-03-12 04:21

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.10827v1</id>\n    <title>Speaker Verification with Speech-Aware LLMs: Evaluation and Augmentation</title>\n    <updated>2026-03-11T14:34:25Z</updated>\n    <link href='https://arxiv.org/abs/2603.10827v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.10827v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Speech-aware large language models (LLMs) can accept speech inputs, yet their training objectives largely emphasize linguistic content or specific fields such as emotions or the speaker's gender, leaving it unclear whether they encode speaker identity. First, we propose a model-agnostic scoring protocol that produces continuous verification scores for both API-only and open-weight models, using confidence scores or log-likelihood ratios from the Yes/No token probabilities. Using this protocol, we benchmark recent speech-aware LLMs and observe weak speaker discrimination (EERs above 20% on VoxCeleb1). Second, we introduce a lightweight augmentation that equips an LLM with ASV capability by injecting frozen ECAPA-TDNN speaker embeddings through a learned projection and training only LoRA adapters. On TinyLLaMA-1.1B, the resulting ECAPA-LLM achieves 1.03% EER on VoxCeleb1-E, approaching a dedicated speaker verification system while preserving a natural-language interface.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.SD'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-03-11T14:34:25Z</published>\n    <arxiv:comment>3 Tables, 1 Figure, Under review</arxiv:comment>\n    <arxiv:primary_category term='cs.SD'/>\n    <author>\n      <name>Thomas Thebaud</name>\n    </author>\n    <author>\n      <name>Yuzhe Wang</name>\n    </author>\n    <author>\n      <name>Laureano Moro-Velazquez</name>\n    </author>\n    <author>\n      <name>Jesus Villalba-Lopez</name>\n    </author>\n    <author>\n      <name>Najim Dehak</name>\n    </author>\n  </entry>"
}