On robust estimation of the location parameter

Book review: Applied cryptography: Protocols, algorithms, and source code in C

Many problems arising in science and engineering aim to find a function which is the optimal value of a specified functional, Some examples include optimal control, inverse analysis and optimal shape design. Only some of these, regarded as variational problems, can be solved analytically, and the only general technique is to approximate the solution using direct methods. Unfortunately, variational problems are very difficult to solve, and it becomes necessary to innovate in the field of numerical methods in order to overcome the difficulties.

The objective of this PhD Thesis is to develop a conceptual theory of neural networks from
the perspective of functional analysis and variational calculus. Within this formulation, learning means to solve a variational problem by minimizing an objective functional associated to the neural network. The choice of the objective functional depends on the particular application. On the other side, its evaluation might need the integration of functions, ordinary differential equations or partial differential equations.

As it will be shown, neural networks are able to deal with a wide range of applications in
mathematics and physics. More specifically, a variational formulation for the multilayer perceptron provides a direct method for solving variational problems. This includes typical applications such as function regression, pattern recognition or time series prediction, but also new ones such as optimal control, inverse problems and optimal shape design.

This addition of applications causes that a standard neural network is not able to deal with some particular problems, and it needs to be augmented. In this work an extended class of multilayer perceptron is developed which, besides the traditional neuron models and network architectures, includes independent parameters, boundary conditions and lower and upper bounds.

The computational performance of this numerical method is investigated here through the solution of different validation problems with analytical solution. Moreover, a variational formulation for an extended class of multilayer perceptron is applied to several engineering cases within optimal control, inverse problems or optimal shape design. Finally, this work comes with the open source neural networks C++ library Flood, which has been implemented following the functional analysis and calculus of variations theories.

https://www.cimne.com/flood/docs/PhDThesis.pdf

Neural networks for variational problems in engineering

This chapter contains sections titled: Introduction The General Model(s) Constructing Concurrence Schemes The Geometry of Shared Secret Schemes Setting Up Shared Secret Schemes Key Distribution via Shared Secret Schemes Conclusions This chapter contains sections titled: References Bibliography (Shared Secret Schemes)

An Introduction to Shared Secret and/or Shared Control Schemes and Their ApplicationThis work was performed at Sandia National Laboratories and supported by the U.S. Department of Energy under contract number DEAC0476DPOO789.

On data-based modelling techniques for fermentation processes

Phishing is a major concern on the Internet today and many users are falling victim because of criminal's deceitful tactics. Blacklisting is still the most common defence users have against such phishing websites, but is failing to cope with the increasing number. In recent years, researchers have devised modern ways of detecting such websites using machine learning. One such method is to create machine learnt models of URL features to classify whether URLs are phishing. However, there are varying opinions on what the best approach is for features and algorithms. In this paper, the objective is to evaluate the performance of the Random Forest algorithm using a lexical only dataset. The performance is benchmarked against other machine learning algorithms and additionally against those reported in the literature. Initial results from experiments indicate that the Random Forest algorithm performs the best yielding an 86.9% accuracy.

/pdf/random-forest-explorations-for-url-classification-571edizcvv.pdf

Random forest explorations for URL classification

A common procedure in digital postproduction is rotoscoping, the segmentation of independently moving foreground elements from background in a sequence of images. Still often carried our manually, rotoscoping is time-consuming and requires great skill in determining the boundary between foreground and background. Errors lead to a bubbling artefact in the final composited sequence. The industry is interested in automated rotoscoping. Any automatic segmentation method must correctly locate the boundary and be robust given rapid motion and non-static backgrounds. A cellular neural network for segmentation is presented that labels pixels by colour, estimated motion and neighbouring labels. The method is accurate, labour saving and many times faster than manual rotoscoping.

Foreground-background segmentation by cellular neural networks

In this paper, we are concerned with techniques for establishing confidentiality of user-generated content (UGC), shared in centralised and untrusted online social networks (OSNs). We describe how the concepts of secret sharing and steganography can be combined to result in a technique for sending confidential messages, as part of a proposed architecture for a virtual private social network (VPSN). We consider the types of UGC confidentiality threats that the VPSN can mitigate, based on those of a decentralised online social network (DOSN). We also postulate the concept of a virtual distributed online social network (VDOSN) in which a VPSN is established across multiple centralised OSNs.

Confidential Communication Techniques for Virtual Private Social Networks

Intrusion detection is concerned with monitoring and analysing events occurring in a computer system in order to discover potential malicious activity. Data mining, which is part of the procedure of knowledge discovery in databases, is the process of analysing the collected data to find patterns or correlations. As the amount of data collected, store and processed only increases, so does the significance and importance of intrusion detection and data mining. A dataset that has been particularly exposed to research is the dataset used for the Third International Knowledge Discovery and Data Mining Tools competition, KDD99. The KDD99 dataset has been used to identify what data mining techniques relate to certain attack and employed to demonstrate that decision trees are more efficient than the Naive Bayes model when it comes to detecting new attacks. When it comes to detecting network intrusions, the C4.5 algorithm performs better than SVM. The aim of our research is to evaluate and compare the usage of various feature selection and reduction algorithms against publicly available datasets. In this contribution, the focus is on feature selection and reduction algorithms. Three feature selection algorithms, consisting of an attribute evaluator and a test method, have been used. Initial results indicate that the performance of the classifier is unaffected by reducing the number of attributes.

An evaluation of feature selection and reduction algorithms for network IDS data

In this contribution, we propose the use of Cellular Neural Networks as an application for the image segmentation of cinematographic image sequences. The proposed approach is based on a Cellular Neural network cost function that takes into account motion and colour. Cellular Neural Networks are of particular interest for hardware implementation due to the inherent parallelism and initial results using an FPGA simulator are also presented. Key-Words: Cellular neural networks, image segmentation, optical flow, motion, FPGA, Handel-C

D. Tsaptsinos

Papers

Random forest explorations for URL classification

Foreground-background segmentation by cellular neural networks

Confidential Communication Techniques for Virtual Private Social Networks

An evaluation of feature selection and reduction algorithms for network IDS data

Image processing using CNNs and FPGAs: Initial results