Contextualization of ASR with LLM using phonetic retrieval-based augmentation
This work addresses the problem of improving named entity recognition in speech for voice assistant applications, representing an incremental advancement in ASR contextualization.
The paper tackled the challenge of recognizing personal named entities in speech by proposing a retrieval-based method to contextualize a large language model (LLM) for automatic speech recognition (ASR), achieving up to 30.2% relative word error rate reduction and 73.6% relative named entity error rate reduction compared to a baseline without contextualization.
Large language models (LLMs) have shown superb capability of modeling multimodal signals including audio and text, allowing the model to generate spoken or textual response given a speech input. However, it remains a challenge for the model to recognize personal named entities, such as contacts in a phone book, when the input modality is speech. In this work, we start with a speech recognition task and propose a retrieval-based solution to contextualize the LLM: we first let the LLM detect named entities in speech without any context, then use this named entity as a query to retrieve phonetically similar named entities from a personal database and feed them to the LLM, and finally run context-aware LLM decoding. In a voice assistant task, our solution achieved up to 30.2% relative word error rate reduction and 73.6% relative named entity error rate reduction compared to a baseline system without contextualization. Notably, our solution by design avoids prompting the LLM with the full named entity database, making it highly efficient and applicable to large named entity databases.