Improving Speech Recognition for Indic Languages using Language Model
This addresses speech recognition accuracy for Indic languages, showing incremental improvements through LM integration.
The study tackled improving speech recognition for Indic languages by applying language models to ASR outputs, resulting in an average CER decrease of over 28% and WER decrease of about 36%.
We study the effect of applying a language model (LM) on the output of Automatic Speech Recognition (ASR) systems for Indic languages. We fine-tune wav2vec $2.0$ models for $18$ Indic languages and adjust the results with language models trained on text derived from a variety of sources. Our findings demonstrate that the average Character Error Rate (CER) decreases by over $28$ \% and the average Word Error Rate (WER) decreases by about $36$ \% after decoding with LM. We show that a large LM may not provide a substantial improvement as compared to a diverse one. We also demonstrate that high quality transcriptions can be obtained on domain-specific data without retraining the ASR model and show results on biomedical domain.