Fixing Errors of the Google Voice Recognizer through Phonetic Distance Metrics

TL;DR

This paper introduces a phonetic distance-based algorithm using Levenshtein distance to improve the accuracy of Google's Spanish speech recognizer, especially for domain-specific phrases, demonstrating significant error reduction.

Contribution

The proposed algorithm applies phonetic Levenshtein distance for domain-specific error correction in speech recognition without modifying the core recognizer.

Findings

Significant error reduction using phonetic distance correction.

Algorithm is general and adaptable to different domains.

Complexity is linear with transcript and phrase set sizes.

Abstract

Speech recognition systems for the Spanish language, such as Google's, produce errors quite frequently when used in applications of a specific domain. These errors mostly occur when recognizing words new to the recognizer's language model or ad hoc to the domain. This article presents an algorithm that uses Levenshtein distance on phonemes to reduce the speech recognizer's errors. The preliminary results show that it is possible to correct the recognizer's errors significantly by using this metric and using a dictionary of specific phrases from the domain of the application. Despite being designed for particular domains, the algorithm proposed here is of general application. The phrases that must be recognized can be explicitly defined for each application, without the algorithm having to be modified. It is enough to indicate to the algorithm the set of sentences on which it must work. The algorithm's complexity is $O(tn)$, where $t$ is the number of words in the transcript to be corrected, and $n$ is the number of phrases specific to the domain.