Tag and correct: high precision post-editing approach to correction of speech recognition errors

TL;DR

The paper proposes a post-editing approach to correcting speech recognition errors using a neural sequence tagger to propose corrections and a separate corrector module to apply them, aiming for architecture-agnostic deployment. It emphasizes high-precision control to avoid introducing new errors, which is critical for production environments. The approach reportedly achieves performance comparable to prior methods while requiring substantially fewer training resources, making it suitable for industrial settings with strict latency and training-time constraints. Overall, the work highlights practical benefits of resource-efficient, reliable ASR error correction in real-world deployments.

Abstract

This paper presents a new approach to the problem of correcting speech recognition errors by means of post-editing. It consists of using a neural sequence tagger that learns how to correct an ASR (Automatic Speech Recognition) hypothesis word by word and a corrector module that applies corrections returned by the tagger. The proposed solution is applicable to any ASR system, regardless of its architecture, and provides high-precision control over errors being corrected. This is especially crucial in production environments, where avoiding the introduction of new mistakes by the error correction model may be more important than the net gain in overall results. The results show that the performance of the proposed error correction models is comparable with previous approaches while requiring much smaller resources to train, which makes it suitable for industrial applications, where both inference latency and training times are critical factors that limit the use of other techniques.