Bayesian network structures are usually built using only the data and starting from an empty network or from a naïve Bayes structure. Very often, in some domains, like medicine, a prior structure knowledge is already known. This structure can be automatically or manually refined in search for better performance models. In this work, we take Bayesian networks built by specialists and show that minor perturbations to this original network can yield better classifiers with a very small computational cost, while maintaining most of the intended meaning of the original model.

We evaluated a population of 7199 children between 2 and 19 years old to study the relations between the observed demographic and physiological features in the occurrence of a pathological/non-pathological heart condition. The data was collected at the Real Hospital Português, Pernambuco, Brazil. We performed a feature importance study, with the aim of categorizing the most relevant variables, indicative of abnormalities. Results show that second heart sound, weight, heart rate, height and secondary reason for consultation are important features, but not nearly as decisive as the presence of heart murmurs. Quantitatively speaking, systolic murmurs and a hyperphonetic second heart sound increase the odds of having a pathology by a factor of 320 and 6, respectively.

We present an improved decoding algorithm for joint turbo decoding and physical-layer network coding. Instead of decoding the individual (binary) messages separately at the relay, the proposed algorithm, from the superimposed faded signals, yields an XOR estimate of the sent messages. Moreover, we introduce a softening of the XOR values to improve the overall performance. Simulation results show that this simple idea yields gains up to 4.5 dB in a Rayleigh fading channel model when compared to a similar scheme.

ABSTRACT The vision towards the Network of the Future cannot be separated from the fact that today's networks, and networking services are subject to sophisticated and very effective attacks. When these attacks first appeared, spoofing and distributed denial‐of‐service attacks were treated as apocalypse for networking. Now, they are considered moderate damage, whereas more sophisticated and inconspicuous attacks, such as botnets activities, might have greater and far reaching impact. As the Internet is expanding to mobile phones and ‘smart dust’ and as its social coverage is liberalized towards the realization of ubiquitous computing (with communication), the concerns on security and privacy have become deeper and the problems more challenging than ever. Re‐designing the Internet as the Network of the Future is self‐motivating for researchers, and security and privacy cannot be provided again as separate, external, add‐on, solutions. In this paper, we discuss the security and privacy challenges of the Network of the Future and try to delimit the solutions space on the basis of emerging techniques. We also review methods that help the quantification of security and privacy in an effort to provide a more systematic and quantitative treatment of the area in the future. Copyright © 2011 John Wiley & Sons, Ltd.

We focus on the problem of secure distributed storage over multiple untrusted clouds or networks. Our main contribution is a low complexity scheme that relies on erasure coding techniques for achieving prescribed levels of confidentiality and reliability. Using matrices that have no singular square submatrices, we subject the original data to a linear transformation. The resulting coded symbols are then stored in different networks. This scheme allows users with access to a threshold number of networks to reconstruct perfectly the original data, while ensuring that eavesdroppers with access to any number of networks smaller than this threshold are unable to decode any of the original symbols. This holds even if the attackers are able to guess some of the missing symbols. We further quantify the achievable level of security, and analyze the complexity of the proposed scheme.

A thorough understanding of the communications channel between vehicles is essential for realistic modeling of Vehicular Ad Hoc Networks (VANETs) and the development of related technology and applications. The impact of vehicles as obstacles on vehicle-to-vehicle (V2V) communication has been largely neglected in VANET research, especially in simulations. Useful models accounting for vehicles as obstacles must satisfy a number of requirements, most notably accurate positioning, realistic mobility patterns, realistic propagation characteristics, and manageable complexity. We present a model that satisfies all of these requirements. Vehicles are modeled as physical obstacles affecting the V2V communication. The proposed model accounts for vehicles as three-dimensional obstacles and takes into account their impact on the LOS obstruction, received signal power, and the packet reception rate. We utilize two real world highway datasets collected via stereoscopic aerial photography to test our proposed model, and we confirm the importance of modeling the effects of obstructing vehicles through experimental measurements. Our results show considerable obstruction of LOS due to vehicles. By obstructing the LOS, vehicles induce significant attenuation and packet loss. The algorithm behind the proposed model allows for computationally efficient implementation in VANET simulators. It is also shown that by modeling the vehicles as obstacles, significant realism can be added to existing simulators with clear implications on the design of upper layer protocols.

A fundamental understanding of the delay behavior of network coding is key towards its successful application in real-time applications with strict message deadlines. Previous contributions focused mostly on the average decoding delay, which although useful in various scenarios of interest is not sufficient for providing worst-case delay guarantees. To overcome this challenge, we investigate the entire delay distribution of random linear network coding for any field size and arbitrary number of encoded symbols (or generation size). By introducing a Markov chain model we are able to obtain a complete solution for the erasure broadcast channel with two receivers. A comparison with Automatic Repeat reQuest (ARQ) with perfect feedback, round robin scheduling and a class of fountain codes reveals that network coding on GF(24) offers the best delay performance for two receivers. We also conclude that GF(2) induces a heavy tail in the delay distribution, which implies that network coding based on XOR operations although simple to implement bears a relevant cost in terms of worst-case delay. For the case of three receivers, which is mathematically challenging, we propose a brute-force methodology that gives the delay distribution of network coding for small generations and field size up to GF(24).

We consider distributed storage over two untrusted networks, whereby coding is used as a means to achieve a prescribed level of confidentiality. The key idea is to exploit the algebraic structure of the Vandermonde matrix to mix the input blocks, before they are stored in different locations. The proposed scheme ensures that eavesdroppers with access to only one of the networks are unable to decode any symbol even if they are capable of guessing some of the missing blocks. Information-theoretic techniques allow us to quantify the achievable level of confidentiality. Moreover, the proposed approach is shown to offer low complexity and optimal rate.

In this paper we proposed a linearly constrained minimum variance receiver for a space-time coded multicarrier (MC) CDMA system in frequency selective fading channels. It is shown that in the proposed receiver the channel can be blindly estimated as the eigenvector that corresponds to the maximum eigenvalue of an autocorrelation matrix, and then, efficient algorithms for subspace tracking can be used. Computer simulations shows that the proposed channel estimation algorithm achieves performance comparable to traditional algorithms with less computational cost.

Blind adaptive and iterative interference cancellation (IC) receiver structures for direct sequence code division multiple access systems in multipath channels are proposed. A code-constrained constant modulus design criterion based on constrained optimisation techniques and adaptive algorithms for receiver and channel parameter estimation are described for successive IC (SIC) and parallel IC (PIC) detectors and a new hybrid IC (HIC) scheme in scenarios subject to multipath fading. The proposed HIC structure combines the strengths of linear, SIC and PIC receivers and is shown to outperform the conventional linear, SIC and PIC structures. A novel iterative detection approach that generates different cancellation orders and selects the most likely symbol estimate on the basis of the instantaneous minimum constant modulus criterion is also proposed and combined with the new HIC structure to further enhance performance. Simulation results for an uplink scenario assess the algorithms, the proposed blind adaptive IC detectors against existing receivers and evaluate the effects of error propagationof the new cancellation techniques.

This paper revisits the subspace-based channel estimation problem in zero-padded systems and proposes an alternative formulation that leads to a solution to which power techniques are readily applicable. Additionally, a lower-complexity post-DFT implementation that also allows the use of power techniques is derived. Both solutions are shown to be equivalent. Furthermore, links to previous related work are established. Mean square error and bit error rate performance analysis of the estimators through computer simulations reveals that for moderate signal to noise ratio, power techniques with powers as low as m = 3 achieves performance similar to the conventional SVD-based estimator.

In this work, we propose a new blind channel estimator based on correlation-matching techniques for zero padded OFDM systems. The correlation matching approach consists in matching the observation covariance matrix, which is parameterized by the unknown channel vector, with its estimates based on the received observations. The channel estimate, within a complex scalar ambiguity, can be obtained blindly as the solution of a minimization problem. Consistency of the estimator is addressed. Mean square error performance of the proposed estimator and the resulting bit-error rate (BER) performance after zero-forcing equalization are analyzed through computer simulations. The results show the robustness of the proposed technique.

This work presents low complexity blind constrained data-reusing adaptive filtering algorithms based on the minimum variance and constant modulus cost functions. Constrained minimum variance (CMV) and constrained constant modulus (CCM) affine projection type algorithms are developed and investigated in a CDMA interference suppression scenario. Computer simulations are used to analyze the proposed techniques and compare them with existing stochastic gradient (SG) and recursive least-squares (RLS) type techniques. The results show that the new algorithms outperform previously reported SG techniques with small additional computational requirements and achieve a performance very close to RLS algorithms at greatly reduced complexity