A Light-weight contextual spelling correction model for customizing transducer-based speech recognition systems

It’s challenging to customize transducer-based automatic speech recognition (ASR) system with context information which is dynamic and unavailable during model training. In this work, we introduce a light-weight contextual spelling correction model to correct context-related recognition errors in transducer-based ASR systems.  We incorporate the context information into the spelling correction model with a shared context encoder and use a filtering algorithm to handle large-size context lists.
Experiments show that the model improves baseline ASR model performance with about 50\% relative word error rate reduction, which also significantly outperforms the baseline method such as contextual LM biasing. The model also shows excellent performance for out-of-vocabulary terms not seen during training.