Towards interfacing large language models with ASR systems using confidence measures and prompting
This work addresses the challenge of enhancing ASR accuracy for systems with lower performance, though it is incremental in nature.
The paper tackles the problem of correcting ASR transcripts using LLMs by proposing confidence-based filtering methods to avoid introducing errors, resulting in improved performance for less competitive ASR systems.
As large language models (LLMs) grow in parameter size and capabilities, such as interaction through prompting, they open up new ways of interfacing with automatic speech recognition (ASR) systems beyond rescoring n-best lists. This work investigates post-hoc correction of ASR transcripts with LLMs. To avoid introducing errors into likely accurate transcripts, we propose a range of confidence-based filtering methods. Our results indicate that this can improve the performance of less competitive ASR systems.