Unifying Global and Near-Context Biasing in a Single Trie Pass

TL;DR

This paper introduces a simple yet effective method combining NE bias lists and n-gram language models to improve rare word recognition and domain adaptation in end-to-end ASR systems with minimal computational cost.

Contribution

It presents a novel integration of global and near-context biasing techniques into a single trie pass for improved ASR performance.

Findings

Entity recognition improved by up to 32% relative.

Overall WER reduced by up to 12% relative.

Effective across multiple languages and datasets.

Abstract

Despite the success of end-to-end automatic speech recognition (ASR) models, challenges persist in recognizing rare, out-of-vocabulary words - including named entities (NE) - and in adapting to new domains using only text data. This work presents a practical approach to address these challenges through an unexplored combination of an NE bias list and a word-level n-gram language model (LM). This solution balances simplicity and effectiveness, improving entities' recognition while maintaining or even enhancing overall ASR performance. We efficiently integrate this enriched biasing method into a transducer-based ASR system, enabling context adaptation with almost no computational overhead. We present our results on three datasets spanning four languages and compare them to state-of-the-art biasing strategies. We demonstrate that the proposed combination of keyword biasing and n-gram LM improves entity recognition by up to 32% relative and reduces overall WER by up to a 12% relative.