Exploration algorithm for learning of sensorimotor tasks using sampling from a weighted Gaussian Mixture

Shitov, D, Pirogova, E, Lech, M and Wysocki, T 2018, 'Exploration algorithm for learning of sensorimotor tasks using sampling from a weighted Gaussian Mixture', in Julien Epps, Joe Wolfe, John Smith and Caroline Jones (ed.) Proceedings of the 17th Australasian International Speech Science and Technology Conference (SST 2018), Sydney, NSW, Australia, 4 - 7 December 2018, pp. 109-112.


Document type: Conference Paper
Collection: Conference Papers

Title Exploration algorithm for learning of sensorimotor tasks using sampling from a weighted Gaussian Mixture
Author(s) Shitov, D
Pirogova, E
Lech, M
Wysocki, T
Year 2018
Conference name SST 2018
Conference location Sydney, NSW, Australia
Conference dates 4 - 7 December 2018
Proceedings title Proceedings of the 17th Australasian International Speech Science and Technology Conference (SST 2018)
Editor(s) Julien Epps, Joe Wolfe, John Smith and Caroline Jones
Publisher Australasian Speech Science and Technology Association
Place of publication Sydney, Australia
Start page 109
End page 112
Total pages 4
Abstract This study presents a sampling efficient algorithm of a goal-directed exploration for learning complex non-linear sensorimotor mappings. The proposed generic approach uses sampling from weighted Gaussian Mixture Models (GMs) with both positive and negative weights that is shown to be an efficient way of searching in a non-linear space with multiple local minima. The simulations were performed by training the articulatory model to learn five distinct sounds of English vowels: [a], [e], [i], [o], [u]. The results demonstrated that after 400 iterations, the algorithm generated sounds with the competence values above 82% for all 5 vowels.
Subjects Signal Processing
Keyword(s) speech acquisition
speech modelling
articulatory synthesis
Gaussian mixtures
Copyright notice Copyright © 2018 ASSTA
ISSN 2207-1296
Versions
Version Filter Type
Access Statistics: 12 Abstract Views  -  Detailed Statistics
Created: Thu, 31 Jan 2019, 11:26:00 EST by Catalyst Administrator
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us