English

CLAR: CIF-Localized Alignment for Retrieval-Augmented Speech LLM-Based Contextual ASR

Sound 2026-03-27 v1

Abstract

Speech LLM-based ASR often struggles with named entities and long-tail words due to strong internal language-model priors. Retrieval-augmented biasing can help, but its effectiveness depends on accurate hotword localization in full-utterance speech under weak supervision. We propose CLAR, a dual-encoder speech-text retriever that uses Continuous Integrate-and-Fire (CIF) to learn monotonic token-level alignments without timestamps. With length-aware localized matching, CLAR anchors short-entity acoustic cues and reduces representation dilution and attention drift. The retriever is trained with a multi-granularity objective combining global and local segment-level contrastive losses and a CIF quantity constraint. At inference, top-ranked hotwords are injected as contextual prompts for the Speech LLM, improving recognition without shallow fusion. Experiments show that CLAR significantly improves hotword retrieval and reduces both CER and B-WER against strong contextual ASR baselines.

Keywords

Cite

@article{arxiv.2603.25460,
  title  = {CLAR: CIF-Localized Alignment for Retrieval-Augmented Speech LLM-Based Contextual ASR},
  author = {Shangkun Huang and Huan Shen and Wei Zou and Yunzhang Chen},
  journal= {arXiv preprint arXiv:2603.25460},
  year   = {2026}
}

Comments

Submitted to Interspeech 2026

R2 v1 2026-07-01T11:39:17.246Z