RNN-T For Latency Controlled ASR With Improved Beam Search
This work addresses latency constraints in speech recognition applications, but it is incremental as it builds on existing RNN-T methods with specific optimizations.
The authors tackled the problem of latency-controlled automatic speech recognition by adapting RNN Transducers for tunable latency and improving their beam search decoding speed, achieving comparable word error rates and better computational efficiency than a hybrid baseline on an English videos dataset.
Neural transducer-based systems such as RNN Transducers (RNN-T) for automatic speech recognition (ASR) blend the individual components of a traditional hybrid ASR systems (acoustic model, language model, punctuation model, inverse text normalization) into one single model. This greatly simplifies training and inference and hence makes RNN-T a desirable choice for ASR systems. In this work, we investigate use of RNN-T in applications that require a tune-able latency budget during inference time. We also improved the decoding speed of the originally proposed RNN-T beam search algorithm. We evaluated our proposed system on English videos ASR dataset and show that neural RNN-T models can achieve comparable WER and better computational efficiency compared to a well tuned hybrid ASR baseline.