Search results for "neural net"

Estudio de la radiación neta en zonas semiáridas utilizando modelos lineales y neuronales y la sinergia entre GERB y SEVIRI

2012

Las regiones áridas o semiáridas se caracterizan por una distribución irregular de los recursos hídricos, lo que muchas veces constituye una limitación para el desarrollo de una determinada región. La variabilidad hidrológica de estas regiones se debe a la mala distribución espacial y temporal de la lluvia, a la topografía heterogénea y a los cambios de origen antropogénicos que muchas veces conducen a procesos de degradación y de desertificación. Como en estas regiones la evapotranspiración explica una parte significativa de la pérdida de agua hacia la atmósfera, el estudio y modelización de la radiación neta en superficie (Rn), es de suma importancia, una vez que las estimaciones o medici…

Contributions and applications around low resource deep learning modeling

2023

El aprendizaje profundo representa la vanguardia del aprendizaje automático en multitud de aplicaciones. Muchas de estas tareas requieren una gran cantidad de recursos computacionales, lo que limita su adopción en dispositivos integrados. El objetivo principal de esta tesis es estudiar métodos y algoritmos que permiten abordar problemas utilizando aprendizaje profundo con bajos recursos computacionales. Este trabajo también tiene como objetivo presentar aplicaciones de aprendizaje profundo en la industria. La primera contribución es una nueva función de activación para redes de aprendizaje profundo: la función de módulo. Los experimentos muestran que la función de activación propuesta logra…

Fault diagnosis of induction motors broken rotor bars by pattern recognition based on noise cancelation

2014

Current signal monitoring (CSM) can be used as an effective tool for diagnosing broken rotor bars fault in induction motors. In this paper, fault diagnosis and classification based on artificial neural networks (ANNs) is done in two stages. In the first stage, a filter is designed to remove irrelevant fault components (such as noise) of current signals. The coefficients of the filter are obtained by least square (LS) algorithm. Then by extracting suitable time domain features from filter's output, a neural network is trained for fault classification. The output vector of this network is represented in one of four categories that includes healthy mode, a 5 mm crack on a bar, one broken bar, …

A sensitivity analysis on artificial neural networks fracture predictions in sheet metal forming operations

2008

Deep-learning based reconstruction of the shower maximum X max using the water-Cherenkov detectors of the Pierre Auger Observatory

2021

The atmospheric depth of the air shower maximum $X_{\mathrm{max}}$ is an observable commonly used for the determination of the nuclear mass composition of ultra-high energy cosmic rays. Direct measurements of $X_{\mathrm{max}}$ are performed using observations of the longitudinal shower development with fluorescence telescopes. At the same time, several methods have been proposed for an indirect estimation of $X_{\mathrm{max}}$ from the characteristics of the shower particles registered with surface detector arrays. In this paper, we present a deep neural network (DNN) for the estimation of $X_{\mathrm{max}}$. The reconstruction relies on the signals induced by shower particles in the groun…

SingleChannelNet : A model for automatic sleep stage classification with raw single-channel EEG

2022

In diagnosing sleep disorders, sleep stage classification is a very essential yet time-consuming process. Various existing state-of-the-art approaches rely on hand-crafted features and multi-modality polysomnography (PSG) data, where prior knowledge is compulsory and high computation cost can be expected. Besides, it is a big challenge to handle the task with raw single-channel electroencephalogram (EEG). To overcome these shortcomings, this paper proposes an end-to-end framework with a deep neural network, namely SingleChannelNet, for automatic sleep stage classification based on raw single-channel EEG. The proposed model utilizes a 90s epoch as the textual input and employs two multi-conv…

Comparison of Machine Learning Methods in Stochastic Skin Optical Model Inversion

2020

In this study, we compare six different machine learning methods in the inversion of a stochastic model for light propagation in layered media, and use the inverse models to estimate four parameters of the skin from the simulated data: melanin concentration, hemoglobin volume fraction, and thicknesses of epidermis and dermis. The aim of this study is to determine the best methods for stochastic model inversion in order to improve current methods in skin related cancer diagnostics and in the future develop a non-invasive way to measure the physical parameters of the skin based partially on the results of the study. Of the compared methods, which are convolutional neural network, multi-layer …

Alleviating Class Imbalance Problem in Automatic Sleep Stage Classification

2022

For real-world automatic sleep-stage classification tasks, various existing deep learning-based models are biased toward the majority with a high proportion. Because of the unique sleep structure, most of the current polysomnography (PSG) datasets suffer an inherent class imbalance problem (CIP), in which the number of each sleep stage is severely unequal. In this study, we first define the class imbalance factor (CIF) to describe the level of CIP quantitatively. Afterward, we propose two balancing methods to alleviate this problem from the dataset quantity and the relationship between the class distribution and the applied model, respectively. The first one is to employ the data augmentati…

SCOPE-Based Emulators for Fast Generation of Synthetic Canopy Reflectance and Sun-Induced Fluorescence Spectra

2017

Progress in advanced radiative transfer models (RTMs) led to an improved understanding of reflectance (R) and sun-induced chlorophyll fluorescence (SIF) emission throughout the leaf and canopy. Among advanced canopy RTMs that have been recently modified to deliver SIF spectral outputs are the energy balance model SCOPE and the 3D models DART and FLIGHT. The downside of these RTMs is that they are computationally expensive, which makes them impractical in routine processing, such as scene generation and retrieval applications. To bypass their computational burden, a computationally effective technique has been proposed by only using a limited number of model runs, called emulation. The idea …

Application of neural networks in diagnostics of chemical compounds based on their infrared spectra

2017

Abstract The paper presents possibilities of using the so-called „finger-print“ identification method and artificial neural network (ANN) for diagnosis of chemical compounds. The construction of a tool specifically developed for this purpose and the ANN, as well as the required conditions for its proper functioning were described. The identification of chemical compounds was tested in two different ways for proving correctness of the assumptions. First of all, initial studies were carried out with the objective to verify the proper functioning of the developed procedure for IR spectrum interpretation. The second research stage was to find out how the properties of artificial neural networks…