Search results for "back"
showing 10 items of 2078 documents
Tabu and Scatter Search for Artificial Neural Networks
2003
In this paper we address the problem of training multilayer feed-forward neural networks. These networks have been widely used for both prediction and classification in many different areas. Although the most popular method for training these networks is back propagation, other optimization methods such as tabu search or scatter search have been applied to solve this problem. This paper presents a new training algorithm based on the tabu search methodology that incorporates elements for search intensification and diversification by utilizing strategic designs where other previous approaches resort to randomization. Our method considers context and search information, as it is provided by th…
Daily Peak Temperature Forecasting with Elman Neural Networks
2005
This work presents a forecaster based on an Elman artificial neural network trained with resilient backpropagation algorithm for predicting the daily peak temperatures one day ahead. The available time series was recorded at Petrosino (TP), in the west coast of Sicily, Italy and it is composed by temperature (min and max values), the humidity (min and max values) and the rainfall value between January 1st, 1995 and May 14th, 2003. Performances and reliabilities of the proposed model were evaluated by a number of measures, comparing different neural models. Experimental results show very good prediction performances.
Connectionist models of face processing: A survey
1994
Abstract Connectionist models of face recognition, identification, and categorization have appeared recently in several disciplines, including psychology, computer science, and engineering. We present a review of these models with the goal of complementing a recent survey by Samal and Iyengar [Pattern Recognition25, 65–77 (1992)] of nonconnectionist approaches to the problem of the automatic face recognition. We concentrate on models that use linear autoassociative networks, nonlinear autoassociative (or compression) and/or heteroassociative backpropagation networks. One advantage of these models over some nonconnectionist approaches is that analyzable features emerge naturally from image-b…
Two-level branch prediction using neural networks
2003
Dynamic branch prediction in high-performance processors is a specific instance of a general time series prediction problem that occurs in many areas of science. Most branch prediction research focuses on two-level adaptive branch prediction techniques, a very specific solution to the branch prediction problem. An alternative approach is to look to other application areas and fields for novel solutions to the problem. In this paper, we examine the application of neural networks to dynamic branch prediction. We retain the first level history register of conventional two-level predictors and replace the second level PHT with a neural network. Two neural networks are considered: a learning vec…
Robust Neutrino Constraints by Combining Low Redshift Observations with the CMB
2009
We illustrate how recently improved low-redshift cosmological measurements can tighten constraints on neutrino properties. In particular we examine the impact of the assumed cosmological model on the constraints. We first consider the new HST H-0 = 74.2 +/- 3.6 measurement by Riess et al. (2009) and the sigma(8)(Omega(m)/0.25)(0.41) = 0.832 +/- 0.033 constraint from Rozo et al. (2009) derived from the SDSS maxBCG Cluster Catalog. In a ACDM model and when combined with WMAP5 constraints, these low-redshift measurements constrain Sigma m(v) < 0.4 eV at the 95% confidence level. This bound does not relax when allowing for the running of the spectral index or for primordial tensor perturbations…
The clustering of galaxies in the SDSS-III Baryon Oscillation Spectroscopic Survey: Baryon Acoustic Oscillations in the Data Release 10 and 11 galaxy…
2014
We present a one per cent measurement of the cosmic distance scale from the detections of the baryon acoustic oscillations in the clustering of galaxies from the Baryon Oscillation Spectroscopic Survey (BOSS), which is part of the Sloan Digital Sky Survey III (SDSS-III). Our results come from the Data Release 11 (DR11) sample, containing nearly one million galaxies and covering approximately $8\,500$ square degrees and the redshift range $0.2<z<0.7$. We also compare these results with those from the publicly released DR9 and DR10 samples. Assuming a concordance $\Lambda$CDM cosmological model, the DR11 sample covers a volume of 13\,Gpc${}^3$ and is the largest region of the Universe ever su…
Cosmological bounds on neutrino statistics
2018
We consider the phenomenological implications of the violation of the Pauli exclusion principle for neutrinos, focusing on cosmological observables such as the spectrum of Cosmic Microwave Background anisotropies, Baryon Acoustic Oscillations and the primordial abundances of light elements. Neutrinos that behave (at least partly) as bosonic particles have a modified equilibrium distribution function that implies a different influence on the evolution of the Universe that, in the case of massive neutrinos, can not be simply parametrized by a change in the effective number of neutrinos. Our results show that, despite the precision of the available cosmological data, only very weak bounds can …
Non-linear evolution of the cosmic neutrino background
2012
We investigate the non-linear evolution of the relic cosmic neutrino background by running large box-size, high resolution N-body simulations which incorporate cold dark matter (CDM) and neutrinos as independent particle species. Our set of simulations explore the properties of neutrinos in a reference Lambda CDM model with total neutrino masses between 0.05-0.60 eV in cold dark matter haloes of mass 10(11) – 10(15) h(-1) M-circle dot, over a redshift range z = 0 – 2. We compute the halo mass function and show that it is reasonably well fitted by the Sheth-Tormen formula, once the neutrino contribution to the total matter is removed. More importantly, we focus on the CDM and neutrino proper…
Constraints on dark matter annihilation from CMB observations before Planck
2013
We compute the bounds on the dark matter (DM) annihilation cross section using the most recent Cosmic Microwave Background measurements from WMAP9, SPT'11 and ACT'10. We consider DM with mass in the MeV-TeV range annihilating 100% into either an e(+)e(-) or a mu(+)mu(-) pair. We consider a realistic energy deposition model, which includes the dependence on the redshift, DM mass and annihilation channel. We exclude the canonical thermal relic abundance cross section ( = 3 x 10(-26) cm(3)s(-1)) for DM masses below 30 GeV and 15 GeV for the e(+)e(-) and mu(+)mu(-) channels, respectively. A priori, DM annihilating in halos could also modify the reionization history of the Universe at late times…
Cosmological parameters degeneracies and non-Gaussian halo bias
2010
We study the impact of the cosmological parameters uncertainties on the measurements of primordial non-Gaussianity through the large-scale non-Gaussian halo bias effect. While this is not expected to be an issue for the standard Lambda CDM model, it may not be the case for more general models that modify the large-scale shape of the power spectrum. We consider the so-called local non-Gaussianity model, parametrized by the f(NL) non-Gaussianity parameter which is zero for a Gaussian case, and make forecasts on f(NL) from planned surveys, alone and combined with a Planck CMB prior. In particular, we consider EUCLID- and LSST-like surveys and forecast the correlations among f(NL) and the runni…