Publications

Network alignment, in general, seeks to discover the hidden underlying correspondence between nodes across two (or more) networks when given their network structure. However, most existing network alignment methods have added assumptions of additional constraints to guide the alignment, such as having a set of seed node-node correspondences across the networks or the existence of side-information. Instead, we seek to develop a general unsupervised network alignment algorithm that makes no additional assumptions. Recently, network embedding has proven effective in many network analysis tasks, but embeddings of different networks are not aligned. Thus, we present our Deep Adversarial Network Alignment (DANA) framework that first uses deep adversarial learning to discover complex mappings for aligning the embedding distributions of the two networks. Then, using our learned mapping functions, DANA performs an efficient nearest neighbor node alignment. Furthermore, we present an unsupervised heuristic to perform model selection for DANA. We perform experiments on real world datasets to show the effectiveness of our framework for first aligning the graph embedding distributions and then discovering node alignments that outperform existing methods.

There are various real-world applications such as video ads, airport screenings, courtroom trials, and job interviews where deception detection can play a crucial role. Hence, there are immense demands on deception detection in videos. However, videos are inherently complex; moreover, they lack detective labels in many real-world applications, which poses tremendous challenges to traditional deception detection methods. In this paper, we study the problem of deception detection in videos. In particular, we provide a principled way to capture rich information into a coherent model and propose an end-to-end framework DEV to detect DEceptive Videos automatically, which is robust to the small number of training data. Experimental results on real-world videos demonstrate the effectiveness of the proposed framework and further experiments are conducted to understand important factors of deception detection in videos.

In order to improve the market value of the product, the platform enterprise often participates in the development process of supporting product of emerging industry’s platform innovation ecosystem. This paper puts forward a revenue sharing contract between the platform company and the supporting company by creating a collaborative development model of the supporting product in the ecosystem, and this paper studies the platform enterprise investment resource property's (complementary or substitution) impact on the supporting enterprise R&D efforts and the revenue sharing factor and analyzes collaborative development mechanism of supporting product of emerging industry platform innovation ecosystem. The research indicates that when platform enterprise and supporting enterprise's resources are complementary, the supporting enterprise R&D effort level and revenue sharing coefficient increase as the platform company’s investment increases. When platform enterprise and supporting enterprise's resources are substitutive, the supporting enterprise’s R&D effort level and revenue sharing coefficient decrease as the platform company’s investment increases.

Artificial Neural Networks (ANNs) have strong potential in modeling nonlinear systems. This paper presents application of a feedforward neural network which utilizes back-propagation learning algorithm, in the area of modeling a vehicle to pole central collision. Kinematics of a typical mid-size vehicle impacting a rigid pole is reproduced by the means of neural networks approach. Firstly, a network is trained with the appropriate data set (acceleration, velocity, and displacement) and subsequently it is tested and simulated. We also provide a comparison concerning the efficiency and performance of each ANN created in this research. It is judged which of them generates the most satisfactory output in the shortest time.

This paper investigated the problem of control design for a class of stochastic systems with Markovian jump parameters and time-varying delays. For the model under consideration, a passivity-based approach is introduced for designing mode-dependent output feedback controllers with mixed discrete and distributed delays. A Lypunov-Krasovskii function (LKF) is defined to establish new required sufficient conditions for ensuring exponentially mean-square stability and the passivity criteria, simultaneously. Moreover, controller gains are calculated based on a convex optimization method by solving a Linear Matrix Inequality (LMI). Finally, simulation results are provided to illustrate the effectiveness of our approach.

This paper investigates the problem of sampled-data (SD) exponentially synchronization for a class of Markovian neural networks with time-varying delayed signals. Based on the tunable parameter and convex combination computational method, a new approach named flexible terminal approach is proposed to reduce the conservatism of delay-dependent synchronization criteria. The SD subject to stochastic sampling period is introduced to exhibit the general phenomena of reality. Novel exponential synchronization criterion are derived by utilizing uniform Lyapunov-Krasovskii functional and suitable integral inequality. Finally, numerical examples are provided to show the usefulness and advantages of the proposed design procedure.

A mixed H 2 / H ∞ output-feedback control design methodology for vibration reduction of base-isolated building structures modelled in the form of second-order linear systems is presented. Sufficient conditions for the design of a desired control are given in terms of linear matrix inequalities. A controller that guarantees asymptotic stability and a mixed H 2 / H ∞ performance for the closed-loop system of the structure is developed, based on a Lyapunov function. The performance of the controller is evaluated by means of simulations in MATLAB/Simulink.

Sales forecasting is an essential task in retailing. In particular, consumer-oriented markets such as fashion and electronics face uncertain demands, short life cycles and a lack of historical sales data which strengthen the challenges of producing accurate forecasts. This survey paper presents state-of-the-art methods in the sales forecasting research with a focus on fashion and new product forecasting. This study also reviews different strategies to the predictive value of user-generated content and search queries.

In decision making problems, experts need to assess a number of values in a side of reference domain higher than in the other one,that is, experts use unbalanced linguistic values to express their evaluation for problems. In this paper, we propose a new linguistic aggregation operator to deal with unbalanced linguistic values in group decision making. Based on 2-tuple representation model of linguistic values and linguistic hierarchies, we express unbalanced linguistic values. Moreover, we present the unbalanced linguistic ordered weighted geometric operator to aggregate unbalanced linguistic evaluation values. Some properties of the unbalanced linguistic ordered weighted geometric operator are also obtained. An illustration example shows the proposed approach seems more effective for linguistic group decision making problem.

This paper deals with modeling and adaptive output tracking of a transverse flux permanent magnet machine as a nonlinear system with unknown nonlinearities by utilizing high gain observer and radial basis function networks. The proposed model is developed based on computing the permeance between rotor and stator using quasiflux tubes. Based on this model, the techniques of feedback linearization and H ∞ control are used to design an adaptive control law for compensating the unknown nonlinear parts, such as the effect of cogging torque, as a disturbance is decreased onto the rotor angle and angular velocity tracking performances. Finally, the capability of the proposed method in tracking both the angle and the angular velocity is shown in the simulation results.

No abstract is provided for this article.

This paper investigates the problem of observer-based robust control for a class of switched stochastic systems with time-varying delay. Based on the average dwell time method, an exponential stability criterion for switched stochastic delay systems is proposed. Then, performance analysis and observer-based robust controller design for the underlying systems are developed. Finally, a numerical example is presented to illustrate the effectiveness of the proposed approach.

Vehicle crash is a complex behavior to be investigated as a challenging topic in terms of dynamical modeling. On this aim, fuzzy logic can be utilized to analyze the crash dynamics rapidly and simply. In this paper, the experimental data of the frontal crash is recorded using an accelerometer located at the centre of the gravity of the vehicle. The acceleration signal was the raw data from which the collision intensity expressed by the kinetic energy and the jerk were derived. The fuzzy logic model was then developed from the two inputs namely kinetic energy and jerk. The output variable is the crash severity expressed as the dynamic crash. The result shows that the jerk contributes much to the crash than the kinetic energy of the vehicle.

Fairness in machine learning has become a global concern due to the predominance of ML in automated decision-making systems. In comparison to group fairness, individual fairness, which aspires that similar individuals should be treated similarly, has received limited attention due to some challenges. One major challenge is the availability of a proper metric to evaluate individual fairness, especially for probabilistic classifiers. In this study, we propose a framework PCIndFair to assess the individual fairness of probabilistic classifiers. Unlike current individual fairness measures, our framework considers probability distribution rather than the final classification outcome, which is suitable for capturing the dynamic of probabilistic classifiers, e.g., neural networks. We perform extensive experiments on four standard datasets and discuss the practical benefits of the framework. This study can be helpful for machine learning researchers and practitioners flexibly assess their models' individual fairness. The complete code of the framework is publicly available <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> .

This paper considers the sliding mode control problem for a kind of dynamic delay system. First by utilizing Lyapunov stability theory and a linear matrix inequality technique, an observer based on delayed output feedback is constructed. Then, an integral sliding surface is presented to realize the sliding mode control for the system with the more available stability condition. Finally, some numerical simulations are implemented to demonstrate the validity of the proposed control method.

In this paper, a novel strategy for structural vibration control of multi-structure systems is presented. This strategy pays particular attention to mitigating negative interstructure interactions. Moreover, it is based on recent advances in static output–feedback control, which make possible the efficient computation of decentralized velocity-feedback controllers by solving a single-step optimization problem with Linear Matrix Inequality constraints. To illustrate the main ideas, a local velocity-feedback energy-to-peak controller is designed for the seismic protection of a two-building system. This controller is remarkably effective and extremely simple. Moreover, it can also be implemented by a linear passive damper. To assess the effectiveness of the proposed controller, numerical simulations are conducted with positive results.