CopyNE: Better Contextual ASR by Copying Named Entities

TL;DR

CopyNE introduces a copying mechanism for end-to-end ASR systems to accurately transcribe contextual named entities as whole units, significantly improving entity transcription accuracy.

Contribution

The paper proposes CopyNE, a novel method that copies entire named entities from a dictionary during ASR, enhancing transcription completeness and accuracy.

Findings

CopyNE improves entity transcription accuracy over previous methods.

CopyNE achieves notable gains even when integrated with strong ASR models like Whisper.

CopyNE reduces errors by copying entire entities at once, ensuring completeness.

Abstract

End-to-end automatic speech recognition (ASR) systems have made significant progress in general scenarios. However, it remains challenging to transcribe contextual named entities (NEs) in the contextual ASR scenario. Previous approaches have attempted to address this by utilizing the NE dictionary. These approaches treat entities as individual tokens and generate them token-by-token, which may result in incomplete transcriptions of entities. In this paper, we treat entities as indivisible wholes and introduce the idea of copying into ASR. We design a systematic mechanism called CopyNE, which can copy entities from the NE dictionary. By copying all tokens of an entity at once, we can reduce errors during entity transcription, ensuring the completeness of the entity. Experiments demonstrate that CopyNE consistently improves the accuracy of transcribing entities compared to previous approaches. Even when based on the strong Whisper, CopyNE still achieves notable improvements.