BR-ASR: Efficient and Scalable Bias Retrieval Framework for Contextual Biasing ASR in Speech LLM
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916757431648256 |
|---|---|
| author | Gong, Xun Lv, Anqi Wang, Zhiming Zhu, Huijia Qian, Yanmin |
| author_facet | Gong, Xun Lv, Anqi Wang, Zhiming Zhu, Huijia Qian, Yanmin |
| contents | While speech large language models (SpeechLLMs) have advanced standard automatic speech recognition (ASR), contextual biasing for named entities and rare words remains challenging, especially at scale. To address this, we propose BR-ASR: a Bias Retrieval framework for large-scale contextual biasing (up to 200k entries) via two innovations: (1) speech-and-bias contrastive learning to retrieve semantically relevant candidates; (2) dynamic curriculum learning that mitigates homophone confusion which negatively impacts the final performance. The is a general framework that allows seamless integration of the retrieved candidates into diverse ASR systems without fine-tuning. Experiments on LibriSpeech test-clean/-other achieve state-of-the-art (SOTA) biased word error rates (B-WER) of 2.8%/7.1% with 2000 bias words, delivering 45% relative improvement over prior methods. BR-ASR also demonstrates high scalability: when expanding the bias list to 200k where traditional methods generally fail, it induces only 0.3 / 2.9% absolute WER / B-WER degradation with a 99.99% pruning rate and only 20ms latency per query on test-other. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_19179 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | BR-ASR: Efficient and Scalable Bias Retrieval Framework for Contextual Biasing ASR in Speech LLM Gong, Xun Lv, Anqi Wang, Zhiming Zhu, Huijia Qian, Yanmin Sound Audio and Speech Processing Signal Processing While speech large language models (SpeechLLMs) have advanced standard automatic speech recognition (ASR), contextual biasing for named entities and rare words remains challenging, especially at scale. To address this, we propose BR-ASR: a Bias Retrieval framework for large-scale contextual biasing (up to 200k entries) via two innovations: (1) speech-and-bias contrastive learning to retrieve semantically relevant candidates; (2) dynamic curriculum learning that mitigates homophone confusion which negatively impacts the final performance. The is a general framework that allows seamless integration of the retrieved candidates into diverse ASR systems without fine-tuning. Experiments on LibriSpeech test-clean/-other achieve state-of-the-art (SOTA) biased word error rates (B-WER) of 2.8%/7.1% with 2000 bias words, delivering 45% relative improvement over prior methods. BR-ASR also demonstrates high scalability: when expanding the bias list to 200k where traditional methods generally fail, it induces only 0.3 / 2.9% absolute WER / B-WER degradation with a 99.99% pruning rate and only 20ms latency per query on test-other. |
| title | BR-ASR: Efficient and Scalable Bias Retrieval Framework for Contextual Biasing ASR in Speech LLM |
| topic | Sound Audio and Speech Processing Signal Processing |
| url | https://arxiv.org/abs/2505.19179 |