Robustness quantification method for network intrusion detection models

Deep learning techniques have been successfully applied to network intrusion detection tasks, but as in the case of autonomous driving and face recognition, the reliability of the system itself has become a pressing issue. Robustness is a key attribute to determine whether a deep learning system is secure and reliable, and we also choose to explore the security of intrusion detection models from a new perspective of robustness quantification. In this paper, we focus on the intrusion detection model based on long and short-term memory, and use a fine-grained linear approximation method to derive a more accurate robustness bound on the nonlinear activation function with tighter linear constraints. We can use this bound to quantitatively measure the robustness of the detection model and determine whether the model is susceptible to the influence of adversarial samples. In our experiments, we test networks with various structures on the MNIST dataset, and the results show that our proposed method can effectively deduce the robustness bounds of output elements, and has good scalability and applicability.     

Robust pulses for high fidelity non-adiabatic geometric gate operations in an off-resonant three-level system

We propose a method to design pulses in a resonant three-level system to enhance the robustness of non-adiabatic geometric gate operations. By optimizing the shape of the pulse envelope, we show that the gate operations are more robust against frequency detuning than they are with Gaussian and square pulses. Our method provides a way to design pulses that can be employed in a system where robustness against frequency variations or inhomogeneous broadening is required, and may be extended to ensure robustness against other physical imperfections such as intensity fluctuations and random noises.     

Robustness of SDOF elastoplastic structure subjected to double-impulse input under simultaneous uncertainties of yield deformation and stiffness

A series of impulse-type inputs has been extensively used to evaluate the critical response of an elastoplastic structure subjected to diverse types of pulse-type inputs, including a near-fault ground motion. In this paper, we consider the critical double-impulse input for a single-degree-of-freedom elastic–perfectly plastic structure, and study effects of structural uncertainties. When the natural frequency (or, equivalently, the stiffness) is fixed, the critical response of the structure does not necessarily decrease as the yield deformation (or, equivalently, the yield force) increases. As the first contribution, we give through investigation of this non-monotonicity property. Moreover, we present a systematic method for finding the worst-case scenario when the yield deformation and the stiffness of a structure assumed to be uncertain simultaneously. Numerical examples are presented to illustrate that the robustness of a structure does not necessarily improve when the yield deformation and/or the yield force is increased.     

A robust decentralized reputation management system for service selection

Reputation mechanism is a novel approach to automate QoS-aware service selection in service oriented computing. The reputation system collects ratings on QoS that consumers feedback and aggregates them to derive a reputation value, which can in turn assist other consumers in service selection in future. However, current approaches fail to combat the malicious ratings and hence the calculated reputation values can be biased severely or even manipulated. Moreover, the centralized management of rating data restricts its application to large open environment. In this paper, we present a robust decentralized reputation system which can resist various unfair ratings and manipulation behaviours. It can evolve and become more mature against malicious ratings with the system running continuously. At last, we experimentally verify the robustness of the proposed approach through a simulation study.     

A Study of Two-Dimensional Plane Elasticity Finite Elements for Optimal Design

ABSTRACT

Design of structures using a variety of two-dimensional finite elements is considered in this paper. An efficient technique of computing first-order derivatives of pointwise stress constraints for simple and higher-order two-dimensional (membrane) finite elements is presented. Computational aspects of design sensitivity vector calculation, using a semi-analytical method versus traditional methods are presented. Implementation of a fully-stressed design approach to find a suitable initial estimate leads to increased computational efficiency. These aspects of the design procedure are illustrated through analysis of numerical examples. Experience indicates that a suitable mix of low and higher-order elements yields the most efficient and accurate design model.     

Robustness of Pathwise Stability of Semilinear Perturbed Stochastic Evolution Equations

Abstract

The purpose of this paper is to investigate pathwise stability for certain Hilbert space-valued stochastic evolution equations. We are especially interested in the robustness analysis of perturbed stochastic differential equations in infinite dimensions. Sufficient conditions are established to ensure the almost surely stable decay of the given stochastic systems. Lastly, a corollary and corresponding example are studied to illustrate our theory.     

Handling uncertainties in toxicity modelling using a fuzzy filter

A fundamental concern in the Quantitative Structure-Activity Relationship approach to toxicity evaluation is the generalization of the model over a wide range of compounds. The data driven modelling of toxicity, due to the complex and ill-defined nature of eco-toxicological systems, is an uncertain process. The development of a toxicity predicting model without considering uncertainties may produce a model with a low generalization performance. This study presents a novel approach to toxicity modelling that handles the involved uncertainties using a fuzzy filter, and thus improves the generalization capability of the model. The method is illustrated by considering a data set dealing with the fathead minnow (Pimephales promelas) toxicity of 568 organic compounds.     

Robustness for Inhomogeneous Poisson Point Processes

We consider robustness for estimation of parametric inhomogeneous Poisson point processes. We propose an influence functional to measure the effect of contamination on estimates. We also propose an M-estimator as an alternative to maximum likelihood estimator, show its consistency and asymptotic normality.     

Robustness and exponentiality in redundant repairable systems

A redundant system of independent repairable components with general failure time and repair time distributions is considered. It is shown that the system failure time has an expectation dependent only on the mean failure times and mean repair times of the components and not on their distributions. An intuitive argument is provided, based on the time of onset of the quasi-stationary distribution on the set of working states, to show that for highly reliable systems, the system failure time is exponentially distributed to good approximation. The argument also justifies a simple figure of merit for the quality of the exponential approximation. A key ingredient in the analysis is the system relaxation time, an informal review of which is provided.     

Study of robustness based on n-way models in the spectrofluorimetric determination of tetracyclines in milk when quenching exists

An important step in the validation of an analytical procedure is the study of its robustness. In the case of spectrofluorimetric determinations, quenching introduces specific problems which are approached in this paper for the particular case of tetracyclines determination in milk. Quenching can be detected with excitation emission matrices (EEM) signals and a three-way Parallel Factor (PARAFAC) decomposition and modelled by means of a four-way PARAFAC decomposition which reproduces the physical model of this effect.The robustness of the method is evaluated by including changes in seven experimental variables: trichloroacetic acid (TCA) volume solution used in the precipitation of milk proteins, revolutions per minute, time and temperature in the centrifugation step, pH and emission-excitation slit width in the fluorimetric analyte determination and the analyst. The robustness analysis is carried out by means of a Plackett-Burman experimental design as it is suggested by European Decision 2002/657/EC (European Decision (EC) No. 2002/657/EC of 12 August 2002, implementing Council Directive 96/23/EC, concerning the performance of analytical methods and the interpretation of results, Off. J.L 221, 17/8/2002, 8). The analyte concentration will be taken as response in the Plackett-Burman experimental design instead of the signal as it is habitual in these cases. Therefore, a three-way Partial Least Squares (3-PLS) calibration models with EEM signal is needed.When an analogous study is carried out for tetracycline (TC) in the absence of chlorotetracycline (CTC) as interferent, univariate calibration is employed, being able to conclude that in the robustness analysis, different factors are significantly active when quenching exists.