A Light-weight contextual spelling correction model for customizing transducer-based speech recognition systems
This work addresses the problem of improving speech recognition accuracy with dynamic context for users of ASR systems, representing a strong specific gain in a domain-specific area.
The paper tackles the challenge of customizing transducer-based speech recognition systems with dynamic context information by introducing a lightweight contextual spelling correction model, which reduces word error rate by about 50% relative to the baseline and outperforms methods like contextual LM biasing.
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.