← All terms

Word error rate

Also known as: WER

The standard metric for evaluating automatic speech recognition accuracy, calculated as the number of substitutions, deletions, and insertions divided by the total number of words in the reference transcript. Research with DHH users has shown that WER correlates poorly with actual user comprehension and satisfaction with captions, highlighting the broader problem that standard AI evaluation metrics may not reflect disabled users' real-world needs.

Category: evaluation · speech recognition · metrics

Related: Automatic speech recognition · Captioning · Benchmark dataset

Sources