Meta-Learning for improving rare word recognition in end-to-end ASR
This addresses the problem of rare word recognition in speech recognition systems, though it appears incremental with a simple interface.
The authors tackled rare word recognition in end-to-end automatic speech recognition by proposing speech embeddings, modified meta-learning approaches for continuous keyword spotting, and a simple integration method, achieving up to 5% improvement in word error rate.
We propose a new method of generating meaningful embeddings for speech, changes to four commonly used meta learning approaches to enable them to perform keyword spotting in continuous signals and an approach of combining their outcomes into an end-to-end automatic speech recognition system to improve rare word recognition. We verify the functionality of each of our three contributions in two experiments exploring their performance for different amounts of classes (N-way) and examples per class (k-shot) in a few-shot setting. We find that the speech embeddings work well and the changes to the meta learning approaches also clearly enable them to perform continuous signal spotting. Despite the interface between keyword spotting and speech recognition being very simple, we are able to consistently improve word error rate by up to 5%.