The reviewed record of science sign in
Pith

arxiv: 2108.01704 · v1 · pith:DBIV7KCG · submitted 2021-08-03 · eess.AS · cs.SD

Bifocal Neural ASR: Exploiting Keyword Spotting for Inference Optimization

Reviewed by Pithpith:DBIV7KCGopen to challenge →

classification eess.AS cs.SD
keywords architecturebifocalinferencerecognitionrnn-tspeechkeywordnetwork
0
0 comments X
read the original abstract

We present Bifocal RNN-T, a new variant of the Recurrent Neural Network Transducer (RNN-T) architecture designed for improved inference time latency on speech recognition tasks. The architecture enables a dynamic pivot for its runtime compute pathway, namely taking advantage of keyword spotting to select which component of the network to execute for a given audio frame. To accomplish this, we leverage a recurrent cell we call the Bifocal LSTM (BFLSTM), which we detail in the paper. The architecture is compatible with other optimization strategies such as quantization, sparsification, and applying time-reduction layers, making it especially applicable for deployed, real-time speech recognition settings. We present the architecture and report comparative experimental results on voice-assistant speech recognition tasks. Specifically, we show our proposed Bifocal RNN-T can improve inference cost by 29.1% with matching word error rates and only a minor increase in memory size.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.