Abstract
We define a new neighborhood called “γ-Observable” and investigate its use in Vector Quantization tasks. Considering a datum v and a set of n units W i in a Euclidean space, let V i be a point of the segment [VW i] whose position depends on γ a real number between 0 and 1, the γ-Observable neighbors (γ-ON) of v are the units W i for which V i is in the Voronoï region of Wi. i.e. W i is the closest unit to V i. For γ=1, V i merges with W i, all the units are y-ON of v, while for γ=O, V i merges with v, only the closest unit to v is γ-ON of v. The size of the neighborhood decreases from n to 1 While γ goes from 1 to O. For γ lower or equal to 0.5, the γ-ON of v are also its natural neighbors, i.e. their Voronoï regions share a common edge with that of v. We experimentally show that this neighborhood used in VQ gives faster convergence and similar final distortion than the “Neural-Gas” onto synthetic data distributions.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Kohonen T. Self-organization and associative memory. Springer Verlag, 1988.
Ritter H, Martinetz T, Schulten K. Neural computation and self-organizing maps. Addison & Wesley, 1992.
Kohonen T, Mäkisara K, Saramäki T. Phonotopics maps - insightful representation of phonological features for speech recognition. In Proc. of 7th Int. Conf. On Pattern Recognition, Montréal, 1984, pp. 182-185.
Makhoul J, Roucos S, Gish H. Vector quantization in speech coding. Proc. IEEE, 1985, vol. 73, pp.1551-1588.
Nasrabadi N M, Feng Y. Vector quantization of images based upon the Kohonen self-organizing feature maps. In IEEE Int. Conf. On Neural Networks, San Diego, CA, 1988, pp. 1101-1108.
Ritter H, Martinetz T, Schulten K. Topology-conserving maps for learning visuomotor- coordination. Neural Networks, 1988, vol.2, pp. 159-168.
Martinetz T, Berkovitch S, Schulten K. A “neural-gas” network for vector quantization and its application to time-series prediction. In IEEE Trans, on Neural Networks 1993, vol. 4, 4: 558-569.
Fortune S. Voronoï diagrams and Delaunay triangulations. Computing in Euclidean Geometry, D.Z.Du, F.Hwang eds. World Scientific 1992, pp. 193-233.
Okabe A, Boots B, Sugihara K. Spatial tessellations: concepts and applications of Voronoï diagrams. John Wiley, Chichester 1992.
Barber C B, Dobkin D P, Huhdanpaa H. The Quickhull algorithm for convex hulls. ACM Trans. on Mathematical Software, December 1996, vol. 22, 4: 469-483. Url:http://www.geom.umn.edu/locate/qhull.
Aupetit M, Lepetz D, Nemoz-Gaillard M, Couturier P, Massotte P. Réseaux de neurones et traitement de données: la notion de voisinage γ-observable. In Valgo 2001, ACTH, April 2001. Url :http://www.supelec-rennes.fr/acth/Valgo/Valgo_Numero-01-01.html.
Eidenbenz S, Stamm C, Widmayer P. Positioning guards at fixed height above a terrain, an optimum inapproximability result. Proc. of Euro. Symposium on Algorithms, 1998.
De Trémiolles G, Tannhof P, Plougonven B, Demarigny C, Madani K. Visual probe mark inspection, using hardware implementation of artificial neural networks, in VLSI production. Proc. of IWANN1997, Lanzarote, Canary Island, Spain 1997.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2001 Springer-Verlag London Limited
About this paper
Cite this paper
Aupetit, M., Couturier, P., Massotte, P. (2001). Vector Quantization with γ-Observable Neighbors. In: Advances in Self-Organising Maps. Springer, London. https://doi.org/10.1007/978-1-4471-0715-6_30
Download citation
DOI: https://doi.org/10.1007/978-1-4471-0715-6_30
Publisher Name: Springer, London
Print ISBN: 978-1-85233-511-3
Online ISBN: 978-1-4471-0715-6
eBook Packages: Springer Book Archive