NoRefER: a Referenceless Quality Metric for Automatic Speech Recognition via Semi-Supervised Language Model Fine-Tuning with Contrastive Learning

TL;DR

NoRefER is a semi-supervised, contrastive learning-based language model that accurately assesses ASR hypothesis quality without needing reference transcripts, enabling efficient model comparison and error detection.

Contribution

It introduces a novel semi-supervised, contrastive learning approach for referenceless ASR quality evaluation using a multilingual language model.

Findings

High correlation with reference-based metrics

Effective intra-sample hypothesis ranking

Potential for referenceless ASR evaluation

Abstract

This paper introduces NoRefER, a novel referenceless quality metric for automatic speech recognition (ASR) systems. Traditional reference-based metrics for evaluating ASR systems require costly ground-truth transcripts. NoRefER overcomes this limitation by fine-tuning a multilingual language model for pair-wise ranking ASR hypotheses using contrastive learning with Siamese network architecture. The self-supervised NoRefER exploits the known quality relationships between hypotheses from multiple compression levels of an ASR for learning to rank intra-sample hypotheses by quality, which is essential for model comparisons. The semi-supervised version also uses a referenced dataset to improve its inter-sample quality ranking, which is crucial for selecting potentially erroneous samples. The results indicate that NoRefER correlates highly with reference-based metrics and their intra-sample ranks, indicating a high potential for referenceless ASR evaluation or a/b testing.