Towards classifying human phonemes without encodings via spatiotemporal liquid state machines

Alex Frid; Hananel Hazan; Larry Manevitz

doi:10.1109/SWSTE.2014.22

Towards classifying human phonemes without encodings via spatiotemporal liquid state machines

Alex Frid, Hananel Hazan, Larry Manevitz

University of Haifa, Technion - Israel Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Classifying human production of phonemes without additional encoding is accomplished at the level of about 77% using a version of reservoir computing. So far this has been accomplished with: (1) artificial data (2) artificial noise (designed to mimic natural noise) (3) natural human data with artificial noise (4) natural human data with its natural noise and variance albeit for certain phonemes. This mechanism, unlike most other methods is done without any encoding of the signal, and without changing time into space, but instead uses the Liquid State Machine paradigm which is an abstraction of natural cortical arrangements. The data is entered as an analogue signal without any modifications. This means that the methodology is close to natural biological mechanisms.

Original language	American English
Title of host publication	Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014
Publisher	IEEE Computer Society
Pages	63-64
Number of pages	2
ISBN (Print)	9780769551883
DOIs	https://doi.org/10.1109/SWSTE.2014.22
State	Published - 2014
Event	2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014 - Ramat Gan, Israel Duration: 11 Jun 2014 → 12 Jun 2014

Publication series

Name	Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014

Conference

Conference	2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014
Country/Territory	Israel
City	Ramat Gan
Period	11/06/14 → 12/06/14

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy

Keywords

Liquid State Machine
Machine Learning
classification
speech synthesis

All Science Journal Classification (ASJC) codes

Software

Access to Document

10.1109/SWSTE.2014.22

Cite this

Frid, A., Hazan, H., & Manevitz, L. (2014). Towards classifying human phonemes without encodings via spatiotemporal liquid state machines. In Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014 (pp. 63-64). Article 6887543 (Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014). IEEE Computer Society. https://doi.org/10.1109/SWSTE.2014.22

Frid, Alex ; Hazan, Hananel ; Manevitz, Larry. / Towards classifying human phonemes without encodings via spatiotemporal liquid state machines. Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014. IEEE Computer Society, 2014. pp. 63-64 (Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014).

@inproceedings{a4c7802e9d9c4791bcf730abd81e9978,

title = "Towards classifying human phonemes without encodings via spatiotemporal liquid state machines",

abstract = "Classifying human production of phonemes without additional encoding is accomplished at the level of about 77\% using a version of reservoir computing. So far this has been accomplished with: (1) artificial data (2) artificial noise (designed to mimic natural noise) (3) natural human data with artificial noise (4) natural human data with its natural noise and variance albeit for certain phonemes. This mechanism, unlike most other methods is done without any encoding of the signal, and without changing time into space, but instead uses the Liquid State Machine paradigm which is an abstraction of natural cortical arrangements. The data is entered as an analogue signal without any modifications. This means that the methodology is close to natural biological mechanisms.",

keywords = "Liquid State Machine, Machine Learning, classification, speech synthesis",

author = "Alex Frid and Hananel Hazan and Larry Manevitz",

year = "2014",

doi = "10.1109/SWSTE.2014.22",

language = "American English",

isbn = "9780769551883",

series = "Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014",

publisher = "IEEE Computer Society",

pages = "63--64",

booktitle = "Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014",

address = "United States",

note = "2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014 ; Conference date: 11-06-2014 Through 12-06-2014",

}

Frid, A, Hazan, H & Manevitz, L 2014, Towards classifying human phonemes without encodings via spatiotemporal liquid state machines. in Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014., 6887543, Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014, IEEE Computer Society, pp. 63-64, 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014, Ramat Gan, Israel, 11/06/14. https://doi.org/10.1109/SWSTE.2014.22

Towards classifying human phonemes without encodings via spatiotemporal liquid state machines. / Frid, Alex; Hazan, Hananel; Manevitz, Larry.
Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014. IEEE Computer Society, 2014. p. 63-64 6887543 (Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Towards classifying human phonemes without encodings via spatiotemporal liquid state machines

AU - Frid, Alex

AU - Hazan, Hananel

AU - Manevitz, Larry

PY - 2014

Y1 - 2014

N2 - Classifying human production of phonemes without additional encoding is accomplished at the level of about 77% using a version of reservoir computing. So far this has been accomplished with: (1) artificial data (2) artificial noise (designed to mimic natural noise) (3) natural human data with artificial noise (4) natural human data with its natural noise and variance albeit for certain phonemes. This mechanism, unlike most other methods is done without any encoding of the signal, and without changing time into space, but instead uses the Liquid State Machine paradigm which is an abstraction of natural cortical arrangements. The data is entered as an analogue signal without any modifications. This means that the methodology is close to natural biological mechanisms.

AB - Classifying human production of phonemes without additional encoding is accomplished at the level of about 77% using a version of reservoir computing. So far this has been accomplished with: (1) artificial data (2) artificial noise (designed to mimic natural noise) (3) natural human data with artificial noise (4) natural human data with its natural noise and variance albeit for certain phonemes. This mechanism, unlike most other methods is done without any encoding of the signal, and without changing time into space, but instead uses the Liquid State Machine paradigm which is an abstraction of natural cortical arrangements. The data is entered as an analogue signal without any modifications. This means that the methodology is close to natural biological mechanisms.

KW - Liquid State Machine

KW - Machine Learning

KW - classification

KW - speech synthesis

UR - http://www.scopus.com/inward/record.url?scp=84907053829&partnerID=8YFLogxK

U2 - 10.1109/SWSTE.2014.22

DO - 10.1109/SWSTE.2014.22

M3 - Conference contribution

SN - 9780769551883

T3 - Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014

SP - 63

EP - 64

BT - Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014

PB - IEEE Computer Society

T2 - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014

Y2 - 11 June 2014 through 12 June 2014

ER -

Frid A, Hazan H, Manevitz L. Towards classifying human phonemes without encodings via spatiotemporal liquid state machines. In Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014. IEEE Computer Society. 2014. p. 63-64. 6887543. (Proceedings - 2014 IEEE International Conference on Software Science, Technology and Engineering, SWSTE 2014). doi: 10.1109/SWSTE.2014.22

Towards classifying human phonemes without encodings via spatiotemporal liquid state machines

Abstract

Publication series

Conference

UN SDGs

Keywords

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this