Stochastic binary hidden units in a multi-layer perceptron (MLP) network give at least three potential benefits when compared to deterministic MLP networks. (1) They allow to learn one-to-many type of mappings. (2) They can be used in structured prediction problems, where modeling the internal structure of the output is important. (3) Stochasticity has been shown to be an excellent regularizer,...

Post-nonlinear (PNL) independent component analysis (ICA) is a generalisation of ICA where the observations are assumed to have been generated from independent sources by linear mixing followed by component-wise scalar nonlinearities. Most previous PNL ICA algorithms require the post-nonlinearities to be invertible functions. In this paper, we present a variational Bayesian approach to PNL ICA ...

The main objective of Present study is based on meteorological variables to estimate monthly Global Solar Radiation (GSR) on a horizontal surface. Monthly mean of maximum air temperature, relative humidity, sunshine hours and wind speed values between 1974 and 2008 for Tehran city in Iran (35_41N, 51_19E), are used in this study. In order to investigate the effect of each meteorological variabl...

In this study, the classification of remotely sensed data using several classifiers and neural networks is considered. The application was conducted using a test scene containing mainly agricultural areas. The main result obtained in this study is that the application of topological map based neural networks to classify the intensity vectors issued from agricultural classes are more suited than...

This paper reviews various optimization techniques available for training multi-layer perception (MLP) artificial neural networks for compression of images. These optimization techniques can be classified into two categories: Derivative-based and Derivative free optimization. The former is based on the calculation of gradient and includes Gradient Descent, Conjugate gradient, Quasi-Newton, Leve...

This paper proposes a strategy to automatically detect the correspondence between measurements of different sensors using object tracking. In addition the strategy includes the ability to learn new features to facilitate the correspondence computation for future measurements. Therefore first a correlation between objects of different modalities is computed using time synchronous changes of attr...

We introduce context embeddings, dense vectors derived from a language model that represent the left/right context of a word instance, and demonstrate that context embeddings significantly improve the accuracy of our transition based parser. Our model consists of a bidirectional LSTM (BiLSTM) based language model that is pre-trained to predict words in plain text, and a multi-layer perceptron (...

An modification of the Approximate Internal Modelbased Neural Control (AIMNC), using Multi Layer Perceptron (MLP) neural networks is proposed. A necessary condition that the system provides zero steady-state error for a constant reference and constant disturbances is derived. In the proposed control strategy only one neural network, which is the neural model of the plant, needs to be trained of...

This paper combines a conventional method of multivariablesystem identification with a dynamic multi-layer perceptron (MLP) toachieve a constructive method of nonlinear system identification. Theclassof nonlinear systems is assumed to operate nominally around anequilibrium point in the neighborhood of which a linearized model existsto represent the system, although norma...

To analyze large sets of digital micrographs from high-throughput screening studies with constant accuracy, advanced image processing algorithms are necessary. In the literature, systems have been proposed applying modelbased fitting algorithms, morphological operators and artificial neural networks (ANN). Because single approaches show limited performance, we propose a hybrid system that combi...