A high-speed, low-resource ASR back-end based on custom arithmetic.

Xiao Li Jonathan Malkin Jeff A. Bilmes

Published in: IEEE Trans. Speech Audio Process. (2006)

Keyphrases

back end
high speed
user friendly
data types
data management
building blocks
version control
detailed design
resource allocation
domain specific
low power
real time
speech recognition
machine learning
data mining
real world
database systems
neural network
floating point
high speed networks
database