Linguistically Informed Evaluation of Multilingual ASR for African Languages

TL;DR

This paper proposes linguistically-informed metrics, including FER and TER, to better evaluate multilingual ASR models for African languages by revealing phonological and tonal errors that traditional WER overlooks.

Contribution

It introduces a new evaluation framework combining FER and a tone-aware extension to expose linguistically meaningful errors in African language ASR models.

Findings

FER and TER reveal error patterns missed by WER.

Models perform better on segmental features than tones.

Tone errors, especially mid and downstep, are most challenging.

Abstract

Word Error Rate (WER) mischaracterizes ASR models' performance for African languages by combining phonological, tone, and other linguistic errors into a single lexical error. By contrast, Feature Error Rate (FER) has recently attracted attention as a viable metric that reveals linguistically meaningful errors in models' performance. In this paper, we evaluate three speech encoders on two African languages by complementing WER with CER, and FER, and add a tone-aware extension (TER). We show that by computing errors on phonological features, FER and TER reveal linguistically-salient error patterns even when word-level accuracy remains low. Our results reveal that models perform better on segmental features, while tones (especially mid and downstep) remain the most challenging features. Results on Yoruba show a striking differential in metrics, with WER=0.788, CER=0.305, and FER=0.151. Similarly for Uneme (an endangered language absent from pretraining data) a model with near-total WER and 0.461 CER achieves the relatively low FER of 0.267. This indicates model error is often attributable to individual phonetic feature errors, which is obscured by all-or-nothing metrics like WER.