LSTM-Based One-Pass Decoder for Low-Latency Streaming

Current state-of-the-art models based on Long-Short Term Memory (LSTM) networks have been extensively used in ASR to improve performance. However, using LSTMs under a streaming setup is not straightforward due to real-time constraints. In this paper we present a novel streaming decoder that includes...

Full description

Saved in:
Bibliographic Details
Main Authors: Jorge, Javier, Gimenez, Adria, Iranzo-Sanchez, Javier, Silvestre-Cerda, Joan Albert, Civera, Jorge, Sanchis, Albert, Juan, Alfons
Format: Conference Proceeding
Language:English
Subjects:
acoustic modeling
Acoustics
automatic speech recognition
Decoding
language modeling
Real-time systems
Signal processing
Speech processing
Speech recognition
streaming
Task analysis
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Current state-of-the-art models based on Long-Short Term Memory (LSTM) networks have been extensively used in ASR to improve performance. However, using LSTMs under a streaming setup is not straightforward due to real-time constraints. In this paper we present a novel streaming decoder that includes a bidirectional LSTM acoustic model as well as an unidirectional LSTM language model to perform the decoding efficiently while keeping the performance comparable to that of an off-line setup. We perform a one-pass decoding using a sliding window scheme for a bidirectional LSTM acoustic model and an LSTM language model. This has been implemented and assessed under a pure streaming setup, and deployed into our production systems. We report WER and latency figures for the well-known LibriSpeech and TED-LIUM tasks, obtaining competitive WER results with low-latency responses.
ISSN:2379-190X
DOI:10.1109/ICASSP40776.2020.9054267