Revisiting the Relation Between Language Model Perplexity and ASR Word Error Rate for Modern End-to-End Speech Recognition

Researchers revisited the relation between language model perplexity and word error rate for modern end-to-end automatic speech recognition systems, finding that internal language modeling capacity and the use of neural LMs and LLMs challenge the previously assumed linear log-log relation. The study shows that ILM subtraction alters the PPL-WER relation, indicating the decoder's internal LM must be considered when evaluating external LM quality.

arXiv:2607.05612v1 Announce Type: new Abstract: Language model LM perplexity PPL has historically been used as a proxy for automatic speech recognition ASR word error rate WER , with prior work reporting an approximately linear relation in log-log space. Modern end-to-end ASR systems challenge this assumption because they already contain internal language modeling capacity, are often evaluated without external language models, and can now be combined with neural LMs and large language models LLMs through different recognition strategies. This paper revisits the relation between PPL and WER for modern ASR systems. We study whether external LMs still improve current end-to-end ASR systems, whether the PPL-WER relation remains linear in log-log space, how encoder context length affects this relation, and how LLM perplexities fit into the trend observed for standard neural LMs. We further investigate internal language modeling ILM in attention-based encoder-decoder systems and show that ILM subtraction changes the observed PPL-WER relation, indicating that the decoder's internal LM must be considered when interpreting the effect of external LM quality.