Untitled

Ritter, Gerhard X.; Urcid, Gonzalo; Iancu, Laurentiu

doi:10.1023/a:1024773330134

Cited by 71 publications

(4 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These noisy images were generated using Matlab. Remarkably, we got perfect recall: error zero, so our GP-generated AM had better behavior compared to the morphological and the alpha-beta models [15] and [34].…”

Section: Results and Analysismentioning

confidence: 96%

“…If a distorted version of a pattern, denoted as X, is fed to M and the obtained output is exactly Y k , then recalling is considered as perfect. The simplicity of AMs mod-els is a result of a great development, which has been carried out during the last 50 years; for some examples, refer to [6,22,24,14,15,25,23,31]. Most of these models have several limitations: limited storage capacity, difficulty to deal with more than one type of pattern (binary, integer or realvalued), lack of robustness to different kinds of noises (additive, subtractive, mixed, Gaussian, etc.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Evolutionary Associative Memories through Genetic Programming

Villegas-Cortéz

Sossa

Avilés‐Cruz

et al. 2012

Parallel Architectures and Bioinspired Algorithms

View full text Add to dashboard Cite

Recibido el 23 de octubre de 2009; aceptado el 10 de enero de 2010Associative Memories (AMs) are useful devices designed to recall output patterns from input patterns. Each input-output pair forms an association. Thus, AMs store associations among pairs of patterns. An important feature is that since its origins AMs have been manually designed. This way, during the last 50 years about 26 different models and variations have been reported. In this paper, we illustrate how new models of AMs can be automatically generated through Genetic Programming (GP) based methodology. In particular, GP provides a way to successfully facilitate the search for an AM in the form of a computer program. The efficiency of the proposal was conducted by means of two tests based on binary and real-valued patterns. The experimental results show that it is possible to automatically generate AMs that achieve good results for the selected pattern recognition problems. This opens a new research area that allows, for the first time, synthesizing new AMs to solve specific problems.

show abstract

Section: Results and Analysismentioning

confidence: 96%

Section: Introductionmentioning

confidence: 99%

Evolutionary Associative Memories through Genetic Programming

Villegas-Cortéz

Sossa

Avilés‐Cruz

et al. 2012

Parallel Architectures and Bioinspired Algorithms

View full text Add to dashboard Cite

show abstract

“…For instance, the studies in [6][7][8][9][10][11][12] are based on the notion of Strong Lattice Independence (SLI), following the conjecture in [13] that SLI vectors are affine independent vectors and thus its convex hull defines a simplex. Using lattice auto-associative memories (LAAM) [14,15] built from the hyperspectral image data, sets of SLI vectors where induced and used as endmembers. Recent study [10] has shown how to obtain sets of affine independent vectors from the rows and columns of the LAAM constructed using the hyperspectral data.…”

Section: Introductionmentioning

confidence: 99%

Endmember induction by lattice associative memories and multi-objective genetic algorithms

Graña

Veganzones

2012

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

Endmembers are the spectral signatures of the constituent materials of an scene captured with a hyperspectral sensor. Endmember induction algorithms (EIAs) try to extract the endmembers of the scene from the corresponding hyperspectral image. In this article, we benefit from recent theoretical results showing that a set of affine independent vectors can be extracted from the rows and columns of lattice autoassociative memories (LAAM). In the linear mixing model (LMM), endmembers are defined as the vertices of a convex polytope covering the data. Affine independence is a sufficient condition for a set of vectors to be the vertices of a convex polytope, and thus to be considered as endmembers. Our basic procedure is the WM algorithm extracting the endmembers from the dual LAAMs built to store the spectra of the hyperspectral image pixels. The set of endmembers induced by this algorithm defines a convex polytope covering the hyperspectral image data. However, the number of induced endmembers obtained by this procedure is too high for practical purposes, besides they are highly correlated. We apply a Multi-Objective Genetic Algorithm (MOGA) to the optimal selection of the image endmembers. Two fitness functions are used, the residual error of the unmixing process and the size of the set of endmembers. From the MOGA's Pareto front we decide the final set of endmembers by examining the decrease in residual error obtained by increasing the number of endmembers. We propose a faster MOGA where the error fitness function is replaced by a fitness function based on the correlation between endmembers. We compare our process with a state-of-the-art EIA on well known benchmark images.

show abstract

“…Las memorias asociativas morfológicas [42] en lugar de sumar los productos de valores y los pesos sinápticos, toman el máximo o el mínimo de las sumas de los valores y sus correspondientes pesos sinápticos, las propiedades de las redes neuronales morfológicas son drásticamente diferentes de los modelos tradiciones de redes neuronales. Para la recuperación de patrones las memorias asociativas morfológicas aun con ruido erosivo [54] (cuando las componentes del vector son menores o iguales a las componentes originales) o con ruido dilatativo (si sus componentes son mayores o iguales a las componentes originales).…”

Section: Memorias Asociativas Morfológicas De Ritter (1998)unclassified

Análisis de expedientes clínicos que apoyan al diagnóstico de cáncer a partir de memorias asociativas evolutivas

Vizueth¹

View full text Add to dashboard Cite

Cancer disease in all its types is still being studied in order to better understand it, given that its condition is of various forms and there are many factors that can be related to a final diagnosis of whether a person has a certain type of cancer. In this work, we present a first approach for the analysis of breast cancer, based on a well-recognized database of clinical records in the pattern recognition study community. The use of evolutionary associative memories is proposed as an analysis tool from machine learning, which according to the research carried out in the state of the art has not been used so far, and these have shown promising results. Our goal is to provide a new point of view of cancer factors as components of patterns; and analyzing the classification behavior from a known database. It should be noted that it is not intended to reduce the size of the pattern, but rather to shed light on the factors possibly related to the disease.

show abstract

Untitled

Cited by 71 publications

References 35 publications

Evolutionary Associative Memories through Genetic Programming

Evolutionary Associative Memories through Genetic Programming

Endmember induction by lattice associative memories and multi-objective genetic algorithms

Análisis de expedientes clínicos que apoyan al diagnóstico de cáncer a partir de memorias asociativas evolutivas

Contact Info

Product

Resources

About