In this paper, we adapt the classical learning algorithm for feed-forward neural networks when monotonicity is required in the inputoutput mapping. Monotonicity can be imposed by adding of suitable penalization terms to the error function. This yields a computationally efficient algorithm with little overhead compared to back-propagation. This algorithm is used to train neural networks for dela...

In this paper we describe and compare two different methods to reduce the cardinality of the set of candidates nodules, characterized by an high sensitivity ratio, and extracted from PA chest radiographs by a fully automatized method. The methods are a rule based system and a feed-forward neural network trained by back-propagation. Both the systems allow to recognize almost the 021 % of false p...

The Adaptive Solutions CN APS architecture chip is a general purpose neurocomputer chip. It has 64 processors, each with 4 K bytes of local memory, running at 25 megahertz. It is capable of implementing most current neural network algorithms with on chip learning. This paper discusses the implementation of the Back Propagation algorithm on an array of these chips and shows performance figures f...

The formation and propagation of singularities for Boltzmann equation in bounded domains has been an important question in numerical studies as well as in theoretical studies. Consider the nonlinear Boltzmann solution near Maxwellians under in-flow, diffuse, or bounce-back boundary conditions. We demonstrate that discontinuity is created at the non-convex part of the grazing boundary, then prop...

In this paper we describe a new high speed method for the calculation of UHF propagation over irregular terrain which is considered to be a homogeneous dielectric. The method, given that it is rigorous and includes both forward and back scattering, is very fast. The antenna pattern, tilt etc. may be specified arbitrarily. As expected, results compare very well with measurements in the case when...

This study high lights on the subject of weight initialization in back-propagation feed-forward networks. Training data is analyzed and the notion of critical points is introduced for determining the initial weights and the number of hidden units. The proposed method has been applied to arti cial data and the publicly available cancer database. The experimental outcomes indicate that the propos...

A speaker independent bimodal phonetic classification experiment regarding the Italian plosive consonants is described. The phonetic classification scheme is based on a feed forward recurrent back-propagation neural network working on audio and visual information. The speech signal is processed by an auditory model producing spectral-like parameters, while the visual signal is processed by a sp...

Connectionist feed-forward networks, t rained with backpropagat ion, can be used both for nonlinear regression and for (discrete one-of-C ) classification. This paper presents approximate Bayesian meth ods to statistical components of back-propagat ion: choosing a cost funct ion and penalty term (interpreted as a form of prior probability), pruning insignifican t weights, est imat ing the uncer...

Data Envelopment Analysis (DEA) is one of the most widely used methods in the measurement efficiency and productivity of Decision Making Units (DMUs). DEA for a large dataset with many inputs/outputs would require huge computer resources in terms of memory and CPU time. This paper introduces a neural network backpropagation Data Envelopment Analysis. Neural network requirements of computer memo...