Dual graph regularized NMF model for social event detection from Flickr data

In this work, we aim to discover real-world events from Flickr data by devising a three-stage event detection framework. In the first stage, a multimodal fusion (MF) model is designed to deal with the heterogeneous feature modalities possessed by the user-shared data, which is advantageous in computation complexity. In the second stage, a dual graph regularized non-negative matrix factorization (DGNMF) model is proposed to learn compact feature representations. DGNMF incorporates Laplacian regularization terms for the data graph and base graph into the objective, keeping the geometry structures underlying the data samples and dictionary bases simultaneously. In the third stage, hybrid clustering algorithms are applied seamlessly to discover event clusters. Extensive experiments conducted on the real-world dataset reveal the MF-DGNMF-based approaches outperform the baselines.     

Efficient distance-based representative skyline computation in 2D space

Representative skyline computation is a fundamental issue in database area, which has attracted much attention in recent years. A notable definition of representative skyline is the distance-based representative skyline (DBRS). Given an integer k, a DBRS includes k representative skyline points that aims at minimizing the maximal distance between a non-representative skyline point and its nearest representative. In the 2D space, the state-of-the-art algorithm to compute the DBRS is based on dynamic programming (DP) which takes O(k m ²) time complexity, where m is the number of skyline points. Clearly, such a DP-based algorithm cannot be used for handling large scale datasets due to the quadratic time cost. To overcome this problem, in this paper, we propose a new approximate algorithm called ARS, and a new exact algorithm named PSRS, based on a carefully-designed parametric search technique. We show that the ARS algorithm can guarantee a solution that is at most ?? larger than the optimal solution. The proposed ARS and PSRS algorithms run in O(klog2mlog(T/??)) and O(k ² log3m) time respectively, where T is no more than the maximal distance between any two skyline points. We also propose an improved exact algorithm, called PSRS+, based on an effective lower and upper bounding technique. We conduct extensive experimental studies over both synthetic and real-world datasets, and the results demonstrate the efficiency and effectiveness of the proposed algorithms.     

Pricing ad slots with consecutive multi-unit demand

We consider the optimal pricing problem for a model of the rich media advertisement market, that has other related applications. Our model differs from traditional position auctions in that we consider buyers whose demand might be multiple consecutive slots, which is motivated by modeling buyers who may require these to display a large size ad. We study three major pricing mechanisms, the Bayesian pricing model, the maximum revenue market equilibrium model and an envy-free solution model. Under the Bayesian model, we design a polynomial-time computable truthful mechanism that optimizes the revenue. For the market equilibrium paradigm, we find a polynomial-time algorithm to obtain the maximum revenue market equilibrium solution. In the envy-free setting, an optimal solution is presented for the case where the buyers have the same demand for the number of consecutive slots. We present results of a simulation that compares the revenues from the above schemes.     

Platform as a service gateway for the Fog of Things

Internet of Things (IoT), one of the key research topics in recent years, together with concepts from Fog Computing, brings rapid advancements in Smart City, Monitoring Systems, industrial control, transportation and other fields. These applications require a reconfigurable sensor architecture that can span multiple scenarios, devices and use cases that allow storage, networking and computational resources to be efficiently used on the edge of the network. There are a number of platforms and gateway architectures that have been proposed to manage these components and enable application deployment. These approaches lack horizontal integration between multiple providers as well as higher order functionalities like load balancing and clustering. This is partly due to the strongly coupled nature of the deployed applications, a lack of abstraction of device communication layers as well as a lock-in for communication protocols. This limitation is a major obstacle for the development of a protocol agnostic application environment that allows for single application to be migrated and to work with multiple peripheral devices with varying protocols from different local gateways. This research looks at existing platforms and their shortcomings as well as proposes a messaging based modular gateway platform that enables clustering of gateways and the abstraction of peripheral communication protocol details. These novelties allow applications to send and receive messages regardless of their deployment location and destination device protocol, creating a more uniform development environment. Furthermore, it results in a more streamlined application development and testing while providing more efficient use of the gateway’s resources. Our evaluation of a prototype for the system shows the need for the migration of resources and the QoS advantages of such a system. The examined use case scenarios show that clustering proves to be an advantage in certain use cases as well as presenting the deployment of a larger testing and control environment through the platform.     

Dynamic neural network method-based improved PSO and BR algorithms for transient probabilistic analysis of flexible mechanism

To improve the computing efficiency and precision of transient probabilistic analysis of flexible mechanism, dynamic neural network method (DNNM)-based improved particle swarm optimization (PSO)/Bayesian regularization (BR) (called as PSO/BR-DNNM) is proposed based on the developed DNNM with the integration of extremum response surface method (ERSM) and artificial neural network (ANN). The mathematical model of DNNM is established based on ANN on the foundation of investigating ERSM. Aiming at the high nonlinearity and strong coupling characteristics of limit state function of flexible mechanism, accurate weights and thresholds of PSO/BR-DNNM function are discussed by searching initial weights and thresholds based on the improved PSO and training final weights and thresholds by the BR-based training performance function. The probabilistic analysis of two-link flexible robot manipulator (TFRM) was investigated with the proposed method. Reliability degree, distribution characteristics and major factors (section sizes of link-2) of TFRM are obtained, which provides a useful reference for a more effective TFRM design. Through the comparison of three methods (Monte Carlo method, DNNM, PSO/BR-DNNM), it is demonstrated that PSO/BR-DNNM reshapes the probability of flexible mechanism probabilistic analysis and improves the computing efficiency while keeping acceptable computational precision. Moreover, the proposed method offers a useful insight for reliability-based design optimization of flexible mechanism and thereby also enriches the theory and method of mechanical reliability design.     

Development of ergonomic posture recognition technique based on 2D ordinary camera for construction hazard prevention through view-invariant features in 2D skeleton motion

Outdoor tasks operated by construction workers are physically demanding, requiring awkward postures leading to pain, injury, accident, or permanent disability. Ergonomic posture recognition (EPR) technique could be a novel solution for ergonomic hazard monitoring and assessment, yet non-intrusiveness and applicability in complex outdoor environment are always critical considerations for device selection in construction site. Thus, we choose RGB camera to capture skeleton motions, which is non-intrusive for workers compared with wearable sensors. It is also stable and widely used in an outdoor construction site considering various light conditions and complex working areas. This study aims to develop an ergonomic posture recognition technique based on 2D ordinary camera for construction hazard prevention through view-invariant features in 2D skeleton motion. Based on captured 2D skeleton motion samples in the test-run, view-invariant features as classifier inputs were extracted to ensure the learned classifier not sensitive to various camera viewpoints and distances to a worker. Three posture classifiers regarding human back, arms, and legs were employed to ensure three postures to be recognized simultaneously in one video frame. The average accuracies of three classifiers in 5-fold cross validation were as high as 95.0%, 96.5%, and 97.6%, respectively, and the overall accuracies tested by three new activities regarding ergonomic assessment scores captured from different camera heights and viewpoints were 89.2%, 88.3%, and 87.6%, respectively. The developed EPR-aided construction accident auto-prevention technique demonstrated robust accuracy to support on-site postural ergonomic assessment for construction workers’ safety and health assurance.     

On the Uniqueness of User Equilibrium Flow with Speed Limit

This technical note focuses on the link flow uniqueness of user equilibrium (UE) with speed limits. Under a mild assumption on the monotonicity of link travel time function, the UE link flow solutions are well recognized to be unique. However, the incorporation of speed limits in the network has undermined the strict monotonicity of link travel time functions, thus the UE flows on the links with speed limits may not be unique. This note addresses the uniqueness problem with two major contributions. First, a polyhedron defined on links is provided, and it is proven that the UE link flow is unique if and only if the polyhedron only contains one value. Second, two concise methods are proposed to mathematically check whether the polyhedron is a singleton, which can be easily solved and convenient for practical use.     

On Convergence of Volterra Series Expansion of a Class of Nonlinear Systems

A fundamental issue in conducting the analysis and design of a nonlinear system via Volterra series theory is how to ensure the excitation magnitude and/or model parameters will be in the appropriate range such that the nonlinear system has a convergent Volterra series expansion. To this aim, parametric convergence bounds of Volterra series expansion of nonlinear systems described by a NARX model, which can reveal under what excitation magnitude or within what parameter range a given NARX system is able to have a convergent Volterra series expansion subject to any given input signal, are investigated systematically in this paper. The existing bound results often are given as a function of the maximum input magnitude, which could be suitable for single‐tone harmonic inputs but very conservative for complicated inputs (e.g. multi‐tone or arbitrary inputs). In this study, the output response of nonlinear systems is expressed in a closed form, which is not only determined by the input magnitude but also related to the input energy or waveform. These new techniques result in more accurate bound criteria, which are not only functions of model parameters and the maximum input magnitude but also consider a factor reflecting the overall input energy or wave form. This is significant to practical applications, since the same nonlinear system could exhibit chaotic behavior subject to a simple single‐tone input but might not with respect to other different input signals (e.g. multi‐tone inputs) of the same input magnitude. The results provide useful guidance for the application of Volterra series‐based theory and methods from an engineering point of view. The Duffing equation is used as a benchmark example to show the effectiveness of the results.     

Bayesian zero-order TSK fuzzy system modeling

Different from the existing TSK fuzzy system modeling methods, a novel zero-order TSK fuzzy modeling method called Bayesian zero-order TSK fuzzy system (B-ZTSK-FS) is proposed from the perspective of Bayesian inference in this paper. The proposed method B-ZTSK-FS constructs zero-order TSK fuzzy system by using the maximum a posteriori (MAP) framework to maximize the corresponding posteriori probability. First, a joint likelihood model about zero-order TSK fuzzy system is defined to derive a new objective function which can assure that both antecedents and consequents of fuzzy rules rather than only their antecedents of the most existing TSK fuzzy systems become interpretable. The defined likelihood model is composed of three aspects: clustering on the training set for antecedents of fuzzy rules, the least squares (LS) error for consequent parameters of fuzzy rules, and a Dirichlet prior distribution for fuzzy cluster memberships which is considered to not only automatically match the “sum-to-one” constraints on fuzzy cluster memberships, but also make the proposed method B-ZTSK-FS scalable for large-scale datasets by appropriately setting the Dirichlet index. This likelihood model indeed indicates that antecedent and consequent parameters of fuzzy rules can be linguistically interpreted and simultaneously optimized by the proposed method B-ZTSK-FS which is based on the MAP framework with the iterative sampling algorithm, which in fact implies that fuzziness and probability can co-jointly work for TSK fuzzy system modeling in a collaborative rather than repulsive way. Finally, experimental results on 28 synthetic and real-world datasets are reported to demonstrate the effectiveness of the proposed method B-ZTSK-FS in the sense of approximation accuracy, interpretability and scalability.     

Generalized Possibilistic Fuzzy C-Means with novel cluster validity indices for clustering noisy data

A generalized form of Possibilistic Fuzzy C-Means (PFCM) algorithm (GPFCM) is presented for clustering noisy data. A function of distance is used instead of the distance itself to damp noise contributions. It is shown that when the data are highly noisy, GPFCM finds accurate cluster centers but FCM (Fuzzy C-Means), PCM (Possibilistic C-Means), and PFCM algorithms fail. FCM, PCM, and PFCM yield inaccurate cluster centers when clusters are not of the same size or covariance norm is used, whereas GPFCM performs well for both of the cases even when the data are noisy. It is shown that generalized forms of FCM and PCM (GFCM and GPCM) are also more accurate than FCM and PCM. A measure is defined to evaluate performance of the clustering algorithms. It shows that average error of GPFCM and its simplified forms are about 80% smaller than those of FCM, PCM, and PFCM. However, GPFCM demands higher computational costs due to nonlinear updating equations. Three cluster validity indices are introduced to determine number of clusters in clean and noisy datasets. One of them considers compactness of the clusters; the other considers separation of the clusters, and the third one considers both separation and compactness. Performance of these indices is confirmed to be satisfactory using various examples of noisy datasets.