Loading...
Loading...

Go to the content (press return)

UNFOLD: a memory-efficient speech recognizer using on-the-fly WFST composition

Author
Yazdani, R.; Arnau, J.; Gonzalez, A.
Type of activity
Presentation of work at congresses
Name of edition
50th Annual IEEE/ACM International Symposium on Microarchitecture
Date of publication
2017
Presentation's date
2017-10-16
Book of congress proceedings
MICRO-50: the 50th Annual IEEE/ACM International Symposium on Microarchitecture: proceedings: October 14-18, 2017, Cambridge, MA
First page
69
Last page
81
Publisher
Association for Computing Machinery (ACM)
DOI
https://doi.org/10.1145/3123939.3124542 Open in new window
Repository
http://hdl.handle.net/2117/113149 Open in new window
URL
https://dl.acm.org/citation.cfm?id=3124542 Open in new window
Abstract
Accurate, real-time Automatic Speech Recognition (ASR) requires huge memory storage and computational power. The main bottleneck in state-of-the-art ASR systems is the Viterbi search on a Weighted Finite State Transducer (WFST). The WFST is a graph-based model created by composing an Acoustic Model (AM) and a Language Model (LM) offline. Offline composition simplifies the implementation of a speech recognizer as only one WFST has to be searched. However, the size of the composed WFST is huge, ty...
Citation
Yazdani, R., Arnau, J., Gonzalez, A. UNFOLD: a memory-efficient speech recognizer using on-the-fly WFST composition. A: Annual IEEE/ACM International Symposium on Microarchitecture. "MICRO-50: the 50th Annual IEEE/ACM International Symposium on Microarchitecture: proceedings: October 14-18, 2017, Cambridge, MA". Cambridge, MA: Association for Computing Machinery (ACM), 2017, p. 69-81.
Keywords
Special purpose systems, Speech recognition
Group of research
ARCO - Microarchitecture and Compilers

Participants