# BR-ASR: Efficient and Scalable Bias Retrieval Framework for Contextual Biasing ASR in Speech LLM

arXiv: https://arxiv.org/abs/2505.19179 · HTML: https://mlsanity.com/paper/2505.19179/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 May 2025.

## Built on by (most cited first, 11 of 11)

- [IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages](https://mlsanity.com/paper/2606.19157/) (2026)
- [PAC: Pronunciation-Aware Contextualized Large Language Model-based Automatic Speech Recognition](https://mlsanity.com/paper/2509.12647/) (2025)
- [When Helpful Context Leaks: Privacy Risks in Domain-Adapted ASR](https://mlsanity.com/paper/2605.28211/) (2026)
- [Attention2Probability: Attention-Driven Terminology Probability Estimation for Robust Speech-to-Text System](https://mlsanity.com/paper/2508.18701/) (2025)
- [RASST: Retrieval-Augmented Simultaneous Speech Translation](https://mlsanity.com/paper/2601.22777/) (2026)
- [JSPG: Dynamic Dictionary Filtering via Joint Semantic-Pinyin-Glyph Retrieval for Chinese Contextual ASR](https://mlsanity.com/paper/2605.16896/) (2026)
- [Error-Aware TF-IDF Retrieval-Augmented Generation for ASR Error Correction](https://mlsanity.com/paper/2606.24915/) (2026)
- [No Detectable Change in Side-Level WER from Prompt-Level Context: A Preregistered Ablation on a Production Oral-History Corpus](https://mlsanity.com/paper/2608.28875/) (2026)
- [NADI 2026: The Second Multidialectal Arabic Speech Processing Shared Task](https://mlsanity.com/paper/2609.27086/) (2026)
- [In-Context Adaptation of Encoder-Decoder Models in Speech Recognition](https://mlsanity.com/paper/2609.33865/) (2026)
- [LLM-HB: Language-Aware LLM-Guided Hotword Biasing for Code-Switching ASR](https://mlsanity.com/paper/2610.16579/) (2026)

## Abstract (arXiv)

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.