BSPNN: boosted subspace probabilistic neural network for email security

In the modern age of Internet connectivity, advanced information systems have accumulated huge volumes of data. Such fast growing, tremendous amount of data, collected and stored in large databases has far exceeded our human ability to comprehend without proper tools. There has been a great deal of research conducted to explore the potential applications of Machine Learning technologies in Security Informatics. This article studies the Network Security Detection problems in which predictive models are constructed to detect network security breaches such as spamming. Due to overwhelming volume of data, complexity and dynamics of computer networks and evolving cyber threats, current security systems suffer limited performance with low detection accuracy and high number of false alarms. To address such performance issues, a novel Machine Learning algorithm, namely Boosted Subspace Probabilistic Neural Network (BSPNN), has been proposed which combines a Radial Basis Function Neural Network with an innovative diversity-based ensemble learning framework. Extensive empirical analyses suggested that BSPNN achieved high detection accuracy with relatively small computational complexity compared with other conventional detection methods.     

Efficient Multimedia Distribution in Source Constraint Networks

In recent years, the number of peer-to-peer (P2P) applications has increased significantly. One important problem in many P2P applications is how to efficiently disseminate data from a single source to multiple receivers on the Internet. A successful model used for analyzing this problem is a graph consisting of nodes and edges, with a capacity assigned to each edge. In some situations however, it is inconvenient to use this model. To that end, we propose to study the problem of efficient data dissemination in a source constraint network. A source constraint network is modeled as a graph in which, the capacity is associated with a node, rather than an edge. The contributions of this paper include (a) a quantitative data dissemination in any source constraint network, (b) a set of topologies suitable for data dissemination in P2P networks, and (c) an architecture and implementation of a P2P system based on the proposed optimal topologies. We will present the experimental results of our P2P system deployed on PlanetLab nodes demonstrating that our approach achieves near optimal throughput while providing scalability, low delay and bandwidth fairness among peers.     

Unsupervised active contours driven by density distance and local fitting energy with applications to medical image segmentation

This study presents an efficient variational region-based active contour model for segmenting images without priori knowledge about the object or background. In order to handle intensity inhomogeneities and noise, we propose to integrate into the region-based local intensity model a global density distance inspired by the Bhattacharyya flow. The local term based on local information of segmented image allows the model to deal with bias field artifact, which arises in data acquisition processes. The global term, which is based on the density distance between the probability distribution functions of image intensity inside and outside the active contour, provides information for accurate segmentation, keeps the curve from spilling, and addresses noise in the image. Intensive 2D and 3D experiments on many imaging modalities of medical fields such as computed tomography, magnetic resonance imaging, and ultrasound images demonstrate the effectiveness of the model when dealing with images with blurred object boundary, intensity inhomogeneities, and noise.     

Experimental and numerical investigation of the debonding interface between an old concrete and an overlay

The paper focuses on debonding propagation along an interface, notably on the major influence of the interlocking between the two faces of the debonding interface. The aim of the study is to obtain the data necessary for relevant and efficient debonding modelling. The work associates experiment and simulation with the purpose of quantifying the interlocking along the interface. The overlay material investigated was a fibre reinforced mortar (FRM). Direct tension tests of notched FRM specimens were firstly conducted to obtain the tensile strength and the residual normal stress—crack width relationship. Its Young's modulus was determined from compression tests. The substrate-overlay interface was investigated by direct tension tests and flexure tests performed on composite substrate-overlay specimens. The direct tension tests provided the interface tensile strength and the relationship between debonding-opening and residual normal tensile stress. Three point flexural static tests informed on the structural behaviour of the interface. The debonding interface propagation was monitored using a video-microscope with a maximum enlargement of ×175. Using the identified and quantified parameters, modelling of the above mentioned static tests was carried out by the finite elements method using CAST3M code developed in France by CEA (Centre for Atomic Energy). The comparison of modelling and experiment results shows a good coherence and proves the important role of interlocking on the debonding mechanism.     

Design and optimization of electrostatic fluid accelerators

Electrostatic air propulsion is a promising technology with such potential applications as energy-efficient ventilation, air sterilization, cooling of electronics, and dehumidification. The challenges of existing designs include the need to increase air speed, backpressure, energy efficiency, and heat exchange capability. The ultimate goal of this direction of research is to create multi-channel energy efficient ionic pumps. In the described project, a single cell analysis is conducted in this study as a building block of future designs. This paper presents the numerical simulation and experimental results of electrostatic fluid accelerators. This study was conducted for the purpose of optimizing device characteristics through the control of the electric field distribution. Simulations were performed for multiple collector electrode voltage distributions. A method to quantify the change in pump performance between different voltage distributions is presented. The influence of space charge on pump performance is also discussed. A significant improvement of air velocity generated by optimized electrostatic fluid accelerators has been achieved using the proposed approach.     

Automatic ARIMA time series modeling for adaptive I/O prefetching

Inadequate I/O performance remains a major challenge in using high-end computing systems effectively. To address this problem, we present TsModeler, an automatic time series modeling and prediction framework for adaptive I/O prefetching that uses ARIMA time series models to predict the temporal patterns of I/O requests. These online pattern analysis techniques and cutoff indicators for autocorrelation patterns enable multistep online predictions suitable for multiblock prefetching. This work also combines time series predictions with spatial Markov model predictions to determine when, what, and how many blocks to prefetch. Experimental results show reductions in execution time compared to the standard Linux file system across various hardware configurations.     

High-performance lift augmentation dynamic seals for turbine bearing compartments

Conventional bearing shaft seal systems used in gas turbine engines are often limited to a sliding velocity of about 100 m/s, differential pressure of 3 bar, gas temperature of 300°C and a seal life less than 8000 h. Advanced engines will require bearing shaft seal systems to operate up to sliding velocity of 200 m/s, differential pressure of 6 bar, gas temperature of 500°C and seal life in excess of 30?000 hours. For seals operating in these advanced conditions, a design with no rubbing contact will be required to achieve long life and reliability. A good validated approach is the use of a gas lift augmentation seal. The design objective for a seal of this type is to have the faces of the seal seek an equilibrium position to avoid any contact. The gap must be small enough to ensure a minimal air leakage, but it must be large enough to limit power dissipation, due to shear in the gas film, and face deformation by shaft displacement, misalignment and vibration. Dynamic seals for a bearing compartment have the following main functions: provide static and dynamic sealing in order to prevent oil leakage from the bearing oil compartment to the air compartment and consequently no oil smell pollution by the use of bleed air; control air leakage to the bearing oil compartment in order to improve performance of the engine and to reduce oil consumption; reduce volume of the oil tank and lubrication system and hence provide weight reduction; to operate in extreme conditions of temperature and with normal and reverse pressure; and reduce the mean time between overhaul (MTBO) and have a very long life. Techspace Aero and Burgmann have carried out design, development and testing of lift augmentation carbon seals and demonstrated that high life and performance levels of these seals are possible in a gas turbine engine environment.     

Improving Parallel Corpus Quality for Chinese-Vietnamese Statistical Machine Translation

The performance of a machine translation system heavily depends on the quantity and quality of the bilingual language resource.However,getting a parallel corpus,which has a large scale and is of high quality,is a very difficult task especially for low resource languages such as Chinese-Vietnamese.Fortunately,multilingual user generated contents (UGC),such as bilingual movie subtitles,provide us access to automatic construction of the parallel corpus.Although the amount of UGC parallel corpora can be considerable,the original corpus is not suitable for statistical machine translation (SMT) systems.The corpus may contain translation errors,sentence mismatching,free translations,etc.To improve the quality of the bilingual corpus for SMT systems,three filtering methods are proposed:sentence length difference,the semantic of sentence pairs,and machine learning.Experiments are conducted on the Chinese to Vietnamese translation corpus.Experimental results demonstrate that all the three methods effectively improve the corpus quality,and the machine translation performance (BLEU score) can be improved by 1.32.     

Improving Parallel Corpus Quality for Chinese-Vietnamese Statistical Machine Translation

The performance of a machine translation system heavily depends on the quantity and quality of the bilingual language resource.However,getting a parallel corpus,which has a large scale and is of high quality,is a very difficult task especially for low resource languages such as ChineseVietnamese.Fortunately,multilingual user generated contents (UGC),such as bilingual movie subtitles,provide us access to automatic construction of the parallel corpus.Although the amount of UGC parallel corpora can be considerable,the original corpus is not suitable for statistical machine translation (SMT) systems.The corpus may contain translation errors,sentence mismatching,free translations,etc.To improve the quality of the bilingual corpus for SMT systems,three filtering methods are proposed:sentence length difference,the semantic of sentence pairs,and machine learning.Experiments are conducted on the Chinese to Vietnamese translation corpus.Experimental results demonstrate that all the three methods effectively improve the corpus quality,and the machine translation performance (BLEU score) can be improved by 1.32.     

An overview of recommender systems in the healthy food domain

Recently, food recommender systems have received increasing attention due to their relevance for healthy living. Most existing studies on the food domain focus on recommendations that suggest proper food items for individual users on the basis of considering their preferences or health problems. These systems also provide functionalities to keep track of nutritional consumption as well as to persuade users to change their eating behavior in positive ways. Also, group recommendation functionalities are very useful in the food domain, especially when a group of users wants to have a dinner together at home or have a birthday party in a restaurant. Such scenarios create many challenges for food recommender systems since the preferences of all group members have to be taken into account in an adequate fashion. In this paper, we present an overview of recommendation techniques for individuals and groups in the healthy food domain. In addition, we analyze the existing state-of-the-art in food recommender systems and discuss research challenges related to the development of future food recommendation technologies.