Search results for "NEURAL NETWORK"
showing 10 items of 1385 documents
A Mushroom Bodies inspired spiking network for classification and sequence learning
2015
Sequence learning is a complex capability shown by living beings, able to extract information from the environment. Looking into the insect world, there are several examples where the presentation time of specific stimuli is considered to select the proper behavioural response. On the basis of previously developed neural models for sequence learning, inspired by the Drosophila melanogaster, a new formalization of key brain structures involved in the process is here provided. The input classification is performed through resonant neurons, stimulated by the complex dynamics generated in a lattice of recurrent spiking neurons modelling the Mushroom Bodies neuropile in the insect brain. The net…
Classification of Sequences with Deep Artificial Neural Networks: Representation and Architectural Issues
2021
DNA sequences are the basic data type that is processed to perform a generic study of biological data analysis. One key component of the biological analysis is represented by sequence classification, a methodology that is widely used to analyze sequential data of different nature. However, its application to DNA sequences requires a proper representation of such sequences, which is still an open research problem. Machine Learning (ML) methodologies have given a fundamental contribution to the solution of the problem. Among them, recently, also Deep Neural Network (DNN) models have shown strongly encouraging results. In this chapter, we deal with specific classification problems related to t…
Optimal implementation of neural activation functions in programmable logic using fuzzy logic
2006
Abstract This work presents a methodology for implementing neural activation function in programmable logic using tools from fuzzy logic. This methodology will allow implementing these intrinsic non-linear functions using comparators and simple linear modellers, easily implemented in programmable logic. This work is particularized to the case of a hyperbolic tangent, the most common function in neural models, showing the excellent results yielded with the proposed approximation.
Deep learning for agricultural land use classification from Sentinel-2
2020
[ES] En el campo de la teledetección se ha producido recientemente un incremento del uso de técnicas de aprendizaje profundo (deep learning). Estos algoritmos se utilizan con éxito principalmente en la estimación de parámetros y en la clasificación de imágenes. Sin embargo, se han realizado pocos esfuerzos encaminados a su comprensión, lo que lleva a ejecutarlos como si fueran “cajas negras”. Este trabajo pretende evaluar el rendimiento y acercarnos al entendimiento de un algoritmo de aprendizaje profundo, basado en una red recurrente bidireccional de memoria corta a largo plazo (2-BiLSTM), a través de un ejemplo de clasificación de usos de suelo agrícola de la Comunidad Valenciana dentro d…
Corrigendum to “Predicting service request in support centers based on nonlinear dynamics, ARMA modeling and neural networks” [Expert Systems with Ap…
2013
Predicting service request in support centers based on nonlinear dynamics, ARMA modeling and neural networks
2008
In this paper, we present the use of different mathematical models to forecast service requests in support centers (SCs). A successful prediction of service request can help in the efficient management of both human and technological resources that are used to solve these eventualities. A nonlinear analysis of the time series indicates the convenience of nonlinear modeling. Neural models based on the time delay neural network (TDNN) are benchmarked with classical models, such as auto-regressive moving average (ARMA) models. Models achieved high values for the correlation coefficient between the desired signal and that predicted by the models (values between 0.88 and 0.97 were obtained in th…
Condition Assessment of Norwegian Bridge Elements Using Existing Damage Records
2020
The Norwegian Public Roads Administration (NPRA) has recorded bridge element damages in a database for all the bridges it manages since the 1990s. This paper presents a comparison of three methods to establish element condition based on damage records. The methods consist in a non-parametric procedure based on the worst damage registered in the element, linear regression considering also bridge and road characteristics data and classification through an artificial neural network. The methods are assessed using a set of 159 bridges inspected in 2016. The results show that diagnostics of bridge element condition can reach high accuracy by using an artificial neural network classifier and taki…
Development of a Design Procedure for Bending Operations
1999
Springback can be considered as one of the most important shape defect in sheet stamping. Such effect results relevant even when simple bending operations are taken into account. In the paper the authors present a design procedure able to provide the proper value of the punch stroke to be applied in order to compensate for elastic springback. In particular two approaches have been followed: firstly an inverse design technique has been utilized in order to find out the response function governing the investigated phenomenon; furthermore neural network techniques have been applied in order to represent the logical link between the input data and the aimed output, i.e. the proper punch stroke …
Unknown order process emulation
2002
Approaches the emulation problem using feedforward neural networks of single input single output (SISO) processes, applying a backpropagation method with a higher convergence rate. In this kind of application, difficult problems appear when the system's order is a priori unknown. A search through the SISO processes space is proposed, aiming to find a favorable neural emulator over the training examples set.
Identification of parameters of dynamic Preisach model by neural networks
2008
In this paper, an approach that allows to identify the parameters of dynamic Preisach model is presented. The fundamental idea of this method is to identify the parameters of a material by using a neural network trained by a collection of hysteresis curves, whose Preisach model is known. After a brief description of dynamic Preisach Model, the neural network that has been used is introduced. The construction of the training data set is illustrated. Finally, the effectiveness of the method is tested on both numerical as well as experimental data.