Time series association learning
Abstract
An acoustic input is recognized from inferred articulatory movements output by a learned relationship between training acoustic waveforms and articulatory movements. The inferred movements are compared with template patterns prepared from training movements when the relationship was learned to regenerate an acoustic recognition. In a preferred embodiment, the acoustic articulatory relationships are learned by a neural network. Subsequent input acoustic patterns then generate the inferred articulatory movements for use with the templates. Articulatory movement data may be supplemented with characteristic acoustic information, e.g. relative power and high frequency data, to improve template recognition.
- Inventors:
-
- Santa Fe, NM
- Issue Date:
- Research Org.:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- OSTI Identifier:
- 870020
- Patent Number(s):
- 5440661
- Assignee:
- United States of America as represented by United States (Washington, DC)
- Patent Classifications (CPCs):
-
G - PHYSICS G10 - MUSICAL INSTRUMENTS G10L - SPEECH ANALYSIS OR SYNTHESIS
- DOE Contract Number:
- W-7405-ENG-36
- Resource Type:
- Patent
- Country of Publication:
- United States
- Language:
- English
- Subject:
- time; series; association; learning; acoustic; input; recognized; inferred; articulatory; movements; output; learned; relationship; training; waveforms; compared; template; patterns; prepared; regenerate; recognition; preferred; embodiment; relationships; neural; network; subsequent; generate; templates; movement; data; supplemented; characteristic; information; relative; power; frequency; improve; acoustic wave; neural network; preferred embodiment; time series; articulatory movements; neural net; /704/
Citation Formats
Papcun, George J. Time series association learning. United States: N. p., 1995.
Web.
Papcun, George J. Time series association learning. United States.
Papcun, George J. Sun .
"Time series association learning". United States. https://www.osti.gov/servlets/purl/870020.
@article{osti_870020,
title = {Time series association learning},
author = {Papcun, George J},
abstractNote = {An acoustic input is recognized from inferred articulatory movements output by a learned relationship between training acoustic waveforms and articulatory movements. The inferred movements are compared with template patterns prepared from training movements when the relationship was learned to regenerate an acoustic recognition. In a preferred embodiment, the acoustic articulatory relationships are learned by a neural network. Subsequent input acoustic patterns then generate the inferred articulatory movements for use with the templates. Articulatory movement data may be supplemented with characteristic acoustic information, e.g. relative power and high frequency data, to improve template recognition.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Sun Jan 01 00:00:00 EST 1995},
month = {Sun Jan 01 00:00:00 EST 1995}
}
Works referenced in this record:
Learning the hidden structure of speech
journal, April 1988
- Elman, Jeffrey L.; Zipser, David
- The Journal of the Acoustical Society of America, Vol. 83, Issue 4
Phrase speech recognition of large vocabulary using feature in articulatory domain
conference, January 1984
- Shirai, K.; Kobayashi, T.
- ICASSP '84. IEEE International Conference on Acoustics, Speech, and Signal Processing
Towards an articulatory phonology
journal, May 1986
- Ohala, John J.; Browman, Catherine P.; Goldstein, Louis M.
- Phonology Yearbook, Vol. 3