Empirical evidence for ultrametric structure in multi-layer perceptron error surfaces

Combinatorial optimization problems share an interesting property with spin glass systems in that their state spaces can exhibit ultrametric structure. We use sampling methods to analyse the error surfaces of feedforward multi-layer perceptron neural networks learning encoder problems. The third order statistics of these points of attraction are examined and found to be arranged in a highly ultrametric way. This is a unique result for a finite, continuous parameter space. The implications of this result are discussed.

Contributo para a modelação do estudo do comportamento de barragens de aterro com o auxílio de redes neuronais artificiais

O conhecimento do comportamento das barragens de aterro é essencial para garantir o seu funcionamento adequado e para que os objetivos de utilização delineados inicialmente para o respetivo aproveitamento hidráulico possam ser cumpridos. Os fatores de maior relevância num estudo deste tipo de barragens, considerando que apenas estão sob solicitações estáticas, são as pressões de água, registadas em piezómetros, os caudais percolados e os deslocamentos superficiais, geralmente medidos em marcas de nivelamento ou em alvos colocados em peças de centragem forçada. Na presente dissertação pretende-se, com base no conhecimento dos registos dessas medições feitas anteriormente e recorrendo a modelos de inteligência artificial, predizer o valor que se obteria em próximas medições, ajudando assim a decidir qual o melhor procedimento para remediar ou tratar um problema de comportamento relacionado com as variáveis em estudo. Esta dissertação divide-se essencialmente em três partes. Primeiramente, introduzem-se os conceitos relativos à segurança de barragens de aterro, de acordo com o regulamento de segurança adotado em Portugal, dando relevo ao tipo de observação a que são submetidas. Seguidamente, introduz-se o conceito de redes neuronais artificiais e apresenta-se a base de dados, criada com o intuito de uniformizar e melhorar a organização dos valores em estudo das barragens de aterro, que têm sido acompanhadas pelo Laboratório Nacional de Engenharia Civil. Com esta pretende-se facilitar a utilização destes elementos por programas de inteligência artificial. Por último, é feito o enquadramento de um caso de estudo, uma barragem de aterro no Norte de Portugal – barragem de Valtorno-Mourão. Utilizando o Neuroph Studio, os dados relativos à observação desta barragem são aplicados numa rede neuronal artificial, Multi Layer Perceptron Backpropagation Neural Network, permitindo antever comportamentos futuros. Os resultados obtidos são discutidos e perspetivam-se trabalhos para continuar a desenvolver a investigação efetuada.

Gear dynamics monitoring using discrete wavelet transformation and multi-layer perceptron neural networks

This paper presents a multi-stage algorithm for the dynamic condition monitoring of a gear. The algorithm provides information referred to the gear status (fault or normal condition) and estimates the mesh stiffness per shaft revolution in case that any abnormality is detected. In the first stage, the analysis of coefficients generated through discrete wavelet transformation (DWT) is proposed as a fault detection and localization tool. The second stage consists in establishing the mesh stiffness reduction associated with local failures by applying a supervised learning mode and coupled with analytical models. To do this, a multi-layer perceptron neural network has been configured using as input features statistical parameters sensitive to torsional stiffness decrease and derived from wavelet transforms of the response signal. The proposed method is applied to the gear condition monitoring and results show that it can update the mesh dynamic properties of the gear on line.

Development and evaluation of neural network freeway incident detection models using field data

This paper discusses a multi-layer feedforward (MLF) neural network incident detection model that was developed and evaluated using field data. In contrast to published neural network incident detection models which relied on simulated or limited field data for model development and testing, the model described in this paper was trained and tested on a real-world data set of 100 incidents. The model uses speed, flow and occupancy data measured at dual stations, averaged across all lanes and only from time interval t. The off-line performance of the model is reported under both incident and non-incident conditions. The incident detection performance of the model is reported based on a validation-test data set of 40 incidents that were independent of the 60 incidents used for training. The false alarm rates of the model are evaluated based on non-incident data that were collected from a freeway section which was video-taped for a period of 33 days. A comparative evaluation between the neural network model and the incident detection model in operation on Melbourne's freeways is also presented. The results of the comparative performance evaluation clearly demonstrate the substantial improvement in incident detection performance obtained by the neural network model. The paper also presents additional results that demonstrate how improvements in model performance can be achieved using variable decision thresholds. Finally, the model's fault-tolerance under conditions of corrupt or missing data is investigated and the impact of loop detector failure/malfunction on the performance of the trained model is evaluated and discussed. The results presented in this paper provide a comprehensive evaluation of the developed model and confirm that neural network models can provide fast and reliable incident detection on freeways. (C) 1997 Elsevier Science Ltd. All rights reserved.

Moottoripolttoaineiden ominaisuuksien mallintaminen

Työn tavoitteena on selvittää voidaanko neuroverkkoa käyttää mallintamaan ja ennustamaan polttoaineen vaikutusta nykyaikaisen auton päästöihin. Näin pystyttäisiin vähentämään aikaa vievien ja kalliiden koeajojen tarvetta. Työ tehtiin Lappeenrannan teknillisen yliopiston ja Fortum Oy:n yhteistyöprojektissa. Työssä tehtiin kolme erilaista mallia. Ensimmäisenä tehtiin autokohtainen malli, jolla pyrittiin ennustamaan autokohtaista käyttäytymistä. Toiseksi kokeiltiin mallia, jossa automalli oli yhtenä syötteenä. Kolmantena yritettiin kiertää eräitä aineiston ongelmia käyttämällä "sumeutettuja" polttoaineiden koostumuksia. Työssä käytettiin MLP-neuroverkkoa, joka opetettiin backpropagation algoritmilla. Työssä havaittiin ettei käytettävissä olleella aineistolla ja käytetyillä malleilla pystytä riittävällä tarkkuudella mallintamaan polttoaineen vaikutusta päästöihin. Aineiston ongelmia olivat mm. suuret mittausvarianssit, aineiston pieni määrä sekä aineiston soveltumattomuus neuroverkolla mallintamiseen.

Method for automatic detection of wheezing in lung sounds

The present report describes the development of a technique for automatic wheezing recognition in digitally recorded lung sounds. This method is based on the extraction and processing of spectral information from the respiratory cycle and the use of these data for user feedback and automatic recognition. The respiratory cycle is first pre-processed, in order to normalize its spectral information, and its spectrogram is then computed. After this procedure, the spectrogram image is processed by a two-dimensional convolution filter and a half-threshold in order to increase the contrast and isolate its highest amplitude components, respectively. Thus, in order to generate more compressed data to automatic recognition, the spectral projection from the processed spectrogram is computed and stored as an array. The higher magnitude values of the array and its respective spectral values are then located and used as inputs to a multi-layer perceptron artificial neural network, which results an automatic indication about the presence of wheezes. For validation of the methodology, lung sounds recorded from three different repositories were used. The results show that the proposed technique achieves 84.82% accuracy in the detection of wheezing for an isolated respiratory cycle and 92.86% accuracy for the detection of wheezes when detection is carried out using groups of respiratory cycles obtained from the same person. Also, the system presents the original recorded sound and the post-processed spectrogram image for the user to draw his own conclusions from the data.

Wear Particle Classifier System Based on an Artificial Neural Network

This paper describes a method of identifying morphological attributes that classify wear particles in relation to the wear process from which they originate and permit the automatic identification without human expertise. The method is based on the use of Multi Layer Perceptron (MLP) for analysis of specific types of microscopic wear particles. The classification of the wear particles was performed according to their morphological attributes of size and aspect ratio, among others. (C) 2010 Journal of Mechanical Engineering. All rights reserved.

Identification of steel bars immersed in reinforced concrete based on experimental results of eddy current testing and artificial neural network analysis

This paper presents an experimental research on the use of eddy current testing (ECT) and artificial neural networks (ANNs) in order to identify the gauge and position of steel bars immersed in concrete structures. The paper presents details of the ECT probe and concrete specimens constructed for the tests, and a study about the influence of the concrete on the values of measured voltages. After this, new measurements were done with a greater number of specimens, simulating a field condition and the results were used to generate training and validation vectors for multilayer perceptron ANNs. The results show a high percentage of correct identification with respect to both, the gauge of the bar and of the thickness of the concrete cover. © 2013 Copyright Taylor and Francis Group, LLC.

Neural network-based wind vector retrieval from satellite scatterometer data

Obtaining wind vectors over the ocean is important for weather forecasting and ocean modelling. Several satellite systems used operationally by meteorological agencies utilise scatterometers to infer wind vectors over the oceans. In this paper we present the results of using novel neural network based techniques to estimate wind vectors from such data. The problem is partitioned into estimating wind speed and wind direction. Wind speed is modelled using a multi-layer perceptron (MLP) and a sum of squares error function. Wind direction is a periodic variable and a multi-valued function for a given set of inputs; a conventional MLP fails at this task, and so we model the full periodic probability density of direction conditioned on the satellite derived inputs using a Mixture Density Network (MDN) with periodic kernel functions. A committee of the resulting MDNs is shown to improve the results.

Neural network-based wind vector retrieval from satellite scatterometer data

Obtaining wind vectors over the ocean is important for weather forecasting and ocean modelling. Several satellite systems used operationally by meteorological agencies utilise scatterometers to infer wind vectors over the oceans. In this paper we present the results of using novel neural network based techniques to estimate wind vectors from such data. The problem is partitioned into estimating wind speed and wind direction. Wind speed is modelled using a multi-layer perceptron (MLP) and a sum of squares error function. Wind direction is a periodic variable and a multi-valued function for a given set of inputs; a conventional MLP fails at this task, and so we model the full periodic probability density of direction conditioned on the satellite derived inputs using a Mixture Density Network (MDN) with periodic kernel functions. A committee of the resulting MDNs is shown to improve the results.

A Non-Destructive Testing Based on Electromagnetic Measurements and Neural Networks for the Inspection of Concrete Structures

This paper presents a new non-destructive testing (NDT) for reinforced concrete structures, in order to identify the components of their reinforcement. A time varying electromagnetic field is generated close to the structure by electromagnetic devices specially designed for this purpose. The presence of ferromagnetic materials (the steel bars of the reinforcement) immersed in the concrete disturbs the magnetic field at the surface of the structure. These field alterations are detected by sensors coils placed on the concrete surface. Variations in position and cross section (the size) of steel bars immersed in concrete originate slightly different values for the induced voltages at the coils.. The values for the induced voltages were obtained in laboratory tests, and multi-layer perceptron artificial neural networks with Levemberg-Marquardt training algorithm were used to identify the location and size of the bar. Preliminary results can be considered very good.

Traffic flow breakdown prediction using feature reduction through Rough-Neuro fuzzy Networks

The prediction of the traffic behavior could help to make decision about the routing process, as well as enables gains on effectiveness and productivity on the physical distribution. This need motivated the search for technological improvements in the Routing performance in metropolitan areas. The purpose of this paper is to present computational evidences that Artificial Neural Network ANN could be use to predict the traffic behavior in a metropolitan area such So Paulo (around 16 million inhabitants). The proposed methodology involves the application of Rough-Fuzzy Sets to define inference morphology for insertion of the behavior of Dynamic Routing into a structured rule basis, without human expert aid. The dynamics of the traffic parameters are described through membership functions. Rough Sets Theory identifies the attributes that are important, and suggest Fuzzy relations to be inserted on a Rough Neuro Fuzzy Network (RNFN) type Multilayer Perceptron (MLP) and type Radial Basis Function (RBF), in order to get an optimal surface response. To measure the performance of the proposed RNFN, the responses of the unreduced rule basis are compared with the reduced rule one. The results show that by making use of the Feature Reduction through RNFN, it is possible to reduce the need for human expert in the construction of the Fuzzy inference mechanism in such flow process like traffic breakdown. © 2011 IEEE.

Improvement of the GenTHREADER method for genomic fold recognition

Motivation: In order to enhance genome annotation, the fully automatic fold recognition method GenTHREADER has been improved and benchmarked. The previous version of GenTHREADER consisted of a simple neural network which was trained to combine sequence alignment score, length information and energy potentials derived from threading into a single score representing the relationship between two proteins, as designated by CATH. The improved version incorporates PSI-BLAST searches, which have been jumpstarted with structural alignment profiles from FSSP, and now also makes use of PSIPRED predicted secondary structure and bi-directional scoring in order to calculate the final alignment score. Pairwise potentials and solvation potentials are calculated from the given sequence alignment which are then used as inputs to a multi-layer, feed-forward neural network, along with the alignment score, alignment length and sequence length. The neural network has also been expanded to accommodate the secondary structure element alignment (SSEA) score as an extra input and it is now trained to learn the FSSP Z-score as a measurement of similarity between two proteins. Results: The improvements made to GenTHREADER increase the number of remote homologues that can be detected with a low error rate, implying higher reliability of score, whilst also increasing the quality of the models produced. We find that up to five times as many true positives can be detected with low error rate per query. Total MaxSub score is doubled at low false positive rates using the improved method.

Volatility smile extrapolation with an artificial neural network

I use a multi-layer feedforward perceptron, with backpropagation learning implemented via stochastic gradient descent, to extrapolate the volatility smile of Euribor derivatives over low-strikes by training the network on parametric prices.

AN ARTIFICIAL NEURAL NETWORK FOR SONAR TARGET DETECTION

Neural Network has emerged as the topic of the day. The spectrum of its application is as wide as from ECG noise filtering to seismic data analysis and from elementary particle detection to electronic music composition. The focal point of the proposed work is an application of a massively parallel connectionist model network for detection of a sonar target. This task is segmented into: (i) generation of training patterns from sea noise that contains radiated noise of a target, for teaching the network;(ii) selection of suitable network topology and learning algorithm and (iii) training of the network and its subsequent testing where the network detects, in unknown patterns applied to it, the presence of the features it has already learned in. A three-layer perceptron using backpropagation learning is initially subjected to a recursive training with example patterns (derived from sea ambient noise with and without the radiated noise of a target). On every presentation, the error in the output of the network is propagated back and the weights and the bias associated with each neuron in the network are modified in proportion to this error measure. During this iterative process, the network converges and extracts the target features which get encoded into its generalized weights and biases.In every unknown pattern that the converged network subsequently confronts with, it searches for the features already learned and outputs an indication for their presence or absence. This capability for target detection is exhibited by the response of the network to various test patterns presented to it.Three network topologies are tried with two variants of backpropagation learning and a grading of the performance of each combination is subsequently made.