Timezone: »

Global Normalization for Streaming Speech Recognition in a Modular Framework
Ehsan Variani · Ke Wu · Michael D Riley · David Rybach · Matt Shannon · Cyril Allauzen

Tue Nov 29 02:00 PM -- 04:00 PM (PST) @ Hall J #801

We introduce the Globally Normalized Autoregressive Transducer (GNAT) for addressing the label bias problem in streaming speech recognition. Our solution admits a tractable exact computation of the denominator for the sequence-level normalization. Through theoretical and empirical results, we demonstrate that by switching to a globally normalized model, the word error rate gap between streaming and non-streaming speech-recognition models can be greatly reduced (by more than 50% on the Librispeech dataset). This model is developed in a modular framework which encompasses all the common neural speech recognition models. The modularity of this framework enables controlled comparison of modelling choices and creation of new models. A JAX implementation of our models has been open sourced.

Author Information

Ehsan Variani (Google)

I am a Staff Research Scientist in Google. My main research focus is statistical and machine learning and information theory with focus on speech and language recognition.

Ke Wu (Google Inc)
Michael D Riley (Google)
David Rybach (Google)
Matt Shannon (Google)
Cyril Allauzen (Google)

More from the Same Authors