Showing 1 to 18 of 18 matching Articles
Results per page:
Export (CSV)
By
Fuentes, Olac; Solorio, Thamar; Terlevich, Roberto; Terlevich, Elena
Show all (4)
1 Citations
In this paper we present an efficient solution, based on active and instancebased machine learning, to the problem of analyzing galactic spectra, an important problem in modern cosmology. The input to the algorithm is the energy flux received from the galaxy; its expected output is the set of stellar populations and dust abundances that make up the galaxy. Our experiments show very accurate results using both noiseless and noisy spectra, and also that a further improvement in accuracy can be obtained when we incorporate prior knowledge obtained from human experts.
more …
By
Gomez, Juan Carlos; Fuentes, Olac
The hybridization of optimization techniques can exploit the strengths of different approaches and avoid their weaknesses. In this work we present a hybrid optimization algorithm based on the combination of Evolution Strategies (ES) and Locally Weighted Linear Regression (LWLR). In this hybrid a local algorithm (LWLR) proposes a new solution that is used by a global algorithm (ES) to produce new better solutions. This new hybrid is applied in solving an interesting and difficult problem in astronomy, the twodimensional fitting of brightness profiles in galaxy images.
The use of standardized fitting functions is arguably the most powerful method for measuring the largescale features (e.g. brightness distribution) and structure of galaxies, specifying parameters that can provide insight into the formation and evolution of galaxies. Here we employ the hybrid algorithm ES+LWLR to find models that describe the bidimensional brightness profiles for a set of optical galactic images. Models are created using two functions: de Vaucoleurs and exponential, which produce models that are expressed as sets of concentric generalized ellipses that represent the brightness profiles of the images.
The problem can be seen as an optimization problem because we need to minimize the difference between the flux from the model and the flux from the original optical image, following a normalized Euclidean distance. We solved this optimization problem using our hybrid algorithm ES+LWLR. We have obtained results for a set of 100 galaxies, showing that hybrid algorithm is very well suited to solve this problem.
more …
By
Calleja, Jorge; Huerta, Gladis; Fuentes, Olac; Benitez, Antonio; Domínguez, Eduardo López; Medina, Ma. Auxilio
Show all (6)
In this paper we present an experimental study of the performance of six machine learning algorithms applied to morphological galaxy classification. We also address the learning approach from imbalanced data sets, inherent to many realworld applications, such as astronomical data analysis problems. We used two oversampling techniques: SMOTE and Resampling, and we vary the amount of generated instances for classification. Our experimental results show that the learning method Random Forest with Resampling obtain the best results for three, five and seven galaxy types, with a Fmeasure about .99 for all cases.
more …
By
Ramírez Cruz, José Federico; Fuentes, Olac; Romero, Rodrigo; Velasco, Aaron
Show all (4)
We present a hybrid method to produce a velocity model of the Earth’s crust using evolutionary and seismic tomography algorithms. This method takes advantage of the global search ability of an evolution strategy and the quick convergence of an iterative threedimensional seismic tomography technique to generate a model of the Earth’s crustal structure from recorded arrival times of wave fronts produced by controlled sources. The evolution strategy finds a threedimensional velocity model with constant lateral velocity layers that minimizes the root mean square residuals computed by the tomographic algorithm. The model found is provided as the initial search point to a first arrival traveltime seismic tomography algorithm, which then computes the final threedimensional velocity model. The method was tested with a realworld data set from an active source experiment performed in the Potrillo Volcanic Field, in Southern New Mexico. Results show that our hybrid method obtains faster convergence and more accurate results than the conventional methods, and does not require an expertsupplied onedimensional model for the seismic tomography procedure.
more …
By
Gomez, Juan Carlos; Fuentes, Olac; Puerari, Ivanio
Fitting brightness profiles of galaxies in one dimension is frequently done because it suffices for some applications and is simple to implement, but many studies now resort to twodimensional fitting, because many wellresolved, nearby galaxies are often poorly fitted by standard onedimensional models. For the fitting we use a model based on de Vaucoleurs and exponential functions that is represented as a set of concentric generalized ellipses that fit the brightness profile of the image. In the end, we have an artificial image that represents the light distribution in the real image, then we make a comparison between such artificial image and the original to measure how close the model is to the real image. The problem can be seen as an optimization problem because we need to minimize the difference between the original optical image and the model, following a normalized Euclidean distance.
In this work we present a solution to such problem from a point of view of optimization using a hybrid algorithm, based on the combination of Evolution Strategies and the QuasiNewton method. Results presented here show that the hybrid algorithm is very well suited to solve the problem, because it can find the solutions in almost all the cases and with a relatively low cost.
more …
By
Zanella, Vittorio; Fuentes, Olac
2 Citations
Image metamorphosis, commonly known as morphing, is a powerful tool for visual effects that consists of the fluid transformation of one digital image into another. There are many techniques for image metamorphosis, but in all of them there is a need for a person to supply the correspondence between the features in the source image and target image. In this paper we describe a method to perform the metamorphosis of face images in frontal view with uniform illumination automatically, using a generic model of a face and evolution strategies to find the features in both face images.
more …
By
Solorio, Thamar; Fuentes, Olac
We present a new algorithm called Ordered Classification, that is useful for classification problems where only few labeled examples are available but a large test set needs to be classified. In many realworld classification problems, it is expensive and some times unfeasible to acquire a large training set, thus, traditional supervised learning algorithms often perform poorly. In our algorithm, classification is performed by a discriminant approach similar to that of Query By Committee within the active learning setting. The method was applied to the realworld astronomical task of automated prediction of stellar atmospheric parameters, as well as to some benchmark learning problems showing a considerable improvement in classification accuracy over conventional algorithms.
more …
By
Salim, Antonio; Fuentes, Olac; Muñoz, Angélica
This paper describes the development of local visionbased behaviors for the robotic soccer domain. The behaviors, which include finding ball, approaching ball, finding goal, approaching goal, shooting and avoiding, have been designed and implemented using a hierarchical control system. The avoiding behavior was learned using the C4.5 rule induction algorithm, the rest of the behaviors were programmed by hand. The object detection system is able to detect the objects of interest at a frame rate of 17 images per second. We compare three pixel classification techniques; one technique is based on color thresholds, another is based on logical AND operations and the last one is based on the artificial life paradigm. Experimental results obtained with a Pioneer 2DX robot equipped with a single camera, playing on an enclosed soccer field with forward role indicate that the robot operates successfully, scoring goals in 90% of the trials.
more …
By
Calleja, Jorge; Fuentes, Olac
Abstract
In this paper we present an experimental study of the performance of three machine learning algorithms applied to the difficult problem of galaxy classification. We use the Naive Bayes classifier, the ruleinduction algorithm C4.5 and a recently introduced classifier named random forest (RF). We first employ image processing to standardize the images, eliminating the effects of orientation and scale, then perform principal component analysis to reduce the dimensionality of the data, and finally, classify the galaxy images. Our experiments show that RF obtains the best results considering three, five and seven galaxy types.
more …
By
Fuentes, Olac; Solorio, Thamar
We present an optimization algorithm that combines active learning and locallyweighted regression to find extreme points of noisy and complex functions. We apply our algorithm to the problem of interferogram analysis, an important problem in optical engineering that is not solvable using traditional optimization schemes and that has received recent attention in the research community. Experimental results show that our method is faster than others previously presented in the literature and that it is very accurate for the case of noiseless interferograms, as well as for the case of interferograms with two types of noise: white noise and intensity gradients, which are due to slight missalignments in the system.
more …
By
Gomez, Juan Carlos; Fuentes, Olac; Athanassoula, Lia; Bosma, Albert
Show all (4)
Abstract
In this work we present Evolution Strategies (ES) as an efficient optimization method for dynamic modelling of the main interacting group of three galaxies in M81. The M81 group is one of the nearest groups of galaxies; its biggest galaxy, M81, sits in the core of the group together with its two companions M82 and NGC3077. The interaction among these three galaxies is very well defined in an image taken in HI. In this first attempt we use nonselfgravitating simulations for modelling dynamically the group; even with this restriction our method reproduces the density distribution of the three galaxies with great precision. Results presented here show that ES is an excellent method to find an accurate model of groups of interacting galaxies, where a global search for a large number of realvalued parameters needs to be performed.
more …
By
Alvarez, Luis; Fuentes, Olac; Terlevich, Roberto
There is now a huge amount of high quality photometric data available in the literature whose analysis is bound to play a fundamental role in studies of the formation and evolution of structure in the Universe. One important problem that this large amount of data generates is the definition of the best procedure or strategy to achieve the best result with the minimum of computational time.
Here we focus on the optimization of methods to obtain stellar population parameters (ages, proportions, redshift and reddening) from photometric data using evolutionary synthesis models. We pose the problem as an optimization problem and we solve it with Evolution Strategies (ES). We also test a hybrid algorithm combining Evolution Strategies and Locally Weighted Linear Regression (LWLR). The experiments show that the hybrid algorithm achieves greater accuracy, and faster convergence than evolution strategies. On the other hand the performance of ES and ESLWLR is similar when noise is added to the input data.
more …
By
Alvarez, Luis; Fuentes, Olac; Terlevich, Roberto
Abstract
There is now a huge amount of high quality photometric data available in the literature whose analysis is bound to play a fundamental role in studies of the formation and evolution of structure in the Universe. One important problem that this large amount of data generates is the definition of the best procedure or strategy to achieve the best result with the minimum of computational time.
Here we focus on the optimization of methods to obtain stellar population parameters (ages, proportions, redshift and reddening) from photometric data using evolutionary synthesis models. We pose the problem as an optimization problem and we solve it with Evolution Strategies (ES). We also test a hybrid algorithm combining Evolution Strategies and Locally Weighted Linear Regression (LWLR). The experiments show that the hybrid algorithm achieves greater accuracy, and faster convergence than evolution strategies. On the other hand the performance of ES and ESLWLR is similar when noise is added to the input data.
more …
By
Zanella, Vittorio; Fuentes, Olac
Image metamorphosis, commonly known as morphing, is a powerful tool for visual effects that consists of the fluid transformation of one digital image into another. There are many techniques for image metamorphosis, but in all of them there is a need for a person to supply the correspondence between the features in the source image and target image. In this paper we describe a method to perform the metamorphosis of face images in frontal view with uniform illumination automatically, using a generic model of a face and evolution strategies to find the features in both face images.
more …
By
Gomez, Juan Carlos; Fuentes, Olac; Athanassoula, Lia; Bosma, Albert
Show all (4)
1 Citations
In this work we present Evolution Strategies (ES) as an efficient optimization method for dynamic modelling of the main interacting group of three galaxies in M81. The M81 group is one of the nearest groups of galaxies; its biggest galaxy, M81, sits in the core of the group together with its two companions M82 and NGC3077. The interaction among these three galaxies is very well defined in an image taken in HI. In this first attempt we use nonselfgravitating simulations for modelling dynamically the group; even with this restriction our method reproduces the density distribution of the three galaxies with great precision. Results presented here show that ES is an excellent method to find an accurate model of groups of interacting galaxies, where a global search for a large number of realvalued parameters needs to be performed.
more …
By
Calleja, Jorge; Fuentes, Olac
5 Citations
In this paper we present an experimental study of the performance of three machine learning algorithms applied to the difficult problem of galaxy classification. We use the Naive Bayes classifier, the ruleinduction algorithm C4.5 and a recently introduced classifier named random forest (RF). We first employ image processing to standardize the images, eliminating the effects of orientation and scale, then perform principal component analysis to reduce the dimensionality of the data, and finally, classify the galaxy images. Our experiments show that RF obtains the best results considering three, five and seven galaxy types.
more …
By
Ramírez, J. Federico; Fuentes, Olac; Gulati, Ravi K.
4 Citations
In this article we present a method for the automated prediction of stellar atmospheric parameters from spectral indices. This method uses a genetic algorithm (GA) for the selection of relevant spectral indices and prototypical stars and predicts their properties, using the knearest neighbors method (KNN). We have applied the method to predict the effective temperature, surface gravity, metallicity, luminosity class and spectral class of stars from spectral indices. Our experimental results show that the feature selection performed by the genetic algorithm reduces the running time of KNN up to 92%, and the predictive accuracy error up to 35%.
more …
By
Ramírez, J. Federico; Fuentes, Olac
1 Citations
In this paper we present a method that combines evolution strategies (ES) and standard optimization algorithms to solve the problem of fitting line profiles of stellar spectra. This method provides a reliable decomposition and a reduction in computing time over conventional algorithms. Using a stellar spectrum as input, we implemented an evolution strategy to find an approximation of the continuum spectrum and spectral lines. After a few generations, the parameters found by ES are given as starting search point to a standard optimization algorithm, which then finds the correct spectral decomposition. We used Gaussian functions to fit spectral lines and the Planck function to represent the continuum spectrum. Our experimental results present the application of this method to real spectra, showing that they can be approximated very accurately.
more …
