Competitive learning algorithms for vector quantization

Ahalt, Stanley C.; Krishnamurthy, Ashok; Chen, Prakoon; Melton, Douglas E.

doi:10.1016/0893-6080(90)90071-r

Cited by 608 publications

(214 citation statements)

References 14 publications

Supporting

Mentioning

211

Contrasting

Unclassified

Order By: Relevance

“…The LVQ networks were trained by using the Frequency Sensitive Competitive Learning algorithm (FSCL) [1], while the MLP nets were trained by using the Back Propagation algorithm. In order to induce diversity among the experts to be combined, different network parameters were used in the training phase.…”

Section: Resultsmentioning

confidence: 99%

A Weighted Majority Vote Strategy Using Bayesian Networks

Cordella¹,

Stefano

Fontanella

2013

Image Analysis and Processing – ICIAP 2013

View full text Add to dashboard Cite

Abstract. Most of the methods for combining classifiers rely on the assumption that the experts to be combined make uncorrelated errors. Unfortunately, this theoretical assumption is not easy to satisfy in practical cases, thus effecting the performance obtainable by applying any combination strategy. We tried to solve this problem by explicitly modeling the dependencies among the experts through the estimation of the joint probability distributions among the outputs of the classifiers and the true class. In this paper we propose a new weighted majority vote rule, that uses the joint probabilities of each class as weights for combining classifier outputs. A Bayesian Network automatically infers the joint probability distribution for each class. The final decision is made by taking into account both the votes received by each class and the statistical behavior of the classifiers. The experimental results confirmed the effectiveness of the proposed method.

show abstract

Section: Resultsmentioning

confidence: 99%

A Weighted Majority Vote Strategy Using Bayesian Networks

Cordella¹,

Stefano

Fontanella

2013

Image Analysis and Processing – ICIAP 2013

View full text Add to dashboard Cite

show abstract

“…The K-means algorithm corresponds to a Hard Competitive Learning technique (Ahalt et al, 1990), where only the closest prototype to the datum is adapted at a time. To escape from local optima of the energy function, it has been improved by defining a neighborhood function which enables the winner to be adapted and also some of its neighbors.…”

Section: K-means Vector Quantizermentioning

confidence: 99%

“…On one hand, if the energy function is a potential then the convergence of the prototypes obeying their adaptation rule toward a minimum of this energy function is well established, in particular in the stochastic optimization framework (Robbins and Monro, 1951;Albert and Gardner, 1967) with which this paper is concerned. For example, the energy function associated to the K-means algorithm (Mac-Queen, 1967;Ahalt et al, 1990), stochastic version of the LBG algorithm of Linde et al (Linde et al, 1980), is a potential as long as the pdf P is continuous (Kohonen, 1991;Pagès, 1993;Cottrell et al, 1998).…”

Section: Introductionmentioning

confidence: 99%

Concerning the differentiability of the energy function in vector quantization algorithms

Lepetz¹,

Nemoz-Gaillard²,

Aupetit

2007

Neural Networks

View full text Add to dashboard Cite

The adaptation rule for Vector Quantization algorithms, and consequently the convergence of the generated sequence, depends on the existence and properties of a function called the energy function, defined on a topological manifold. Our aim is to investigate the conditions of existence of such a function for a class of algorithms examplified by the initial "K-means" (Mac-Queen, 1967) and Kohonen algorithms (Kohonen, 1982;Kohonen, 1988). The results presented here supplement previous studies, including (Tolat, 1990), (Erwin et al., 1992), (Cottrell et al., 1994), (Pagès, 1993) and (Cottrell et al., 1998). Our work shows that the energy function is not always a potential but at least the uniform limit of a series of potential functions which we call a pseudo-potential. It also shows that a large number of existing vector quantization algorithms developed by the Artificial Neural Networks community fall into this category. The framework we define opens the way to study the convergence of all the corresponding adaptation rules at once, and a theorem gives promising insights in that direction. We also demonstrate that the "K-means" energy function is a pseudo-potential but not a potential in general. Consequently, the energy function associated to the "Neural-Gas" is not a potential in general.

show abstract

“…Acoustic speech signals are represented as sequences of vector quantization ( V Q codes (Ahalt, Krishnamurthy, Chen & Melton, 1990;Gray, 1984;Linde, Buzo & Gray, 1980 describe vector quantization). In section II.C we show an example of how Malcom can be applied to the case where the distribution of articulator positions that produce a code is assumed to be Gaussian.…”

Section: /5/96mentioning

confidence: 99%

Improving on hidden Markov models: An articulatorily constrained, maximum likelihood approach to speech recognition and speech coding

Hogden¹

1996

View full text Add to dashboard Cite

Competitive learning algorithms for vector quantization

Cited by 608 publications

References 14 publications

A Weighted Majority Vote Strategy Using Bayesian Networks

A Weighted Majority Vote Strategy Using Bayesian Networks

Concerning the differentiability of the energy function in vector quantization algorithms

Improving on hidden Markov models: An articulatorily constrained, maximum likelihood approach to speech recognition and speech coding

Contact Info

Product

Resources

About