Search

As one of most successful industrial realizations of Internet of Things, a smart grid is a smart IoT system that deploys widespread smart meters to capture fine-grained data on residential power usage. It suffers diverse privacy attacks, which seriously increases the risk of violating the privacy of customers. Although some solutions have been proposed to address this privacy issue, most of them mainly rely on a trusted party and focus on the sanitization of metering masurements. However, these solutions are vulnerable to advanced attacks. In this paper, we propose a decentralized mechanism for privacy-preserving computation in smart grid called DDP, which leaverages the differential privacy and extends the data sanitization from the value domain to the time domain. Specifically, we inject Laplace noise to the measurements at the end of each customer in a distributed manner, and then use a random permutation algorithm to shuffle the power measurement sequence, thereby enforcing differential privacy after aggregation and preventing the sensitive power usage mode informaton of the customers from being inferred by other parties. Extensive experiments demonstrate that DDP shows an outstanding performance in terms of privacy from the non-intrusive load monitoring (NILM) attacks and utility by using two different error analysis.

There are several kinds of smart devices, such as smartphones, sensors, and smart wearable devices, included in the Human-in-the-Loop (HITL) system, but different devices have their own data processing and programming paradigm. Programmers usually need to design the same data processing logic for different devices by using a different programming model. How to mapping the same code to different devices without any change is an emerging topic in the HITL system. Furthermore, the intelligent data processing for the smart CPS sector is experiencing significant growth in data volume, driven by a large number of smart devices that are anticipated in the near further. All these smart devices are expected to improve the overall HITL system performance marvelously. A large number of devices can also outstandingly increase the data volume, which needs to be processed in real time. How to process large-scale data on a smart device in real time is another challenge. Focused on these challenges, this article proposed a computing device-aware HITL CPS data processing framework, named Barge, aiming to map the regular code to the different hardware without any change. In Barge, a semantic model, an architecture-driven programming model, and a graph partition scheme are included. The semantic model is used to express the user-defined graph algorithms by using the domain-specific language. The architecture-driven programming model will execute the graph algorithms on a different device in parallel. Furthermore, the graph partition scheme will partition the large-scale graphs into suitable partitions by aware of the topology to make the partitioned data suitable for kinds of smart devices. We believe that our work would open a wide range of opportunities to improve the performance of large-scale graph processing for HITL systems.

Fingerprint provides an extremely convenient way of identification for a wide range of real-life applications owing to its universality, uniqueness, collectability, and invariance. However, digitized fingerprints may reveal the privacy of individuals. Differential privacy is a promising privacy-preserving solution that is enforced by injecting random noise into preserved objects, such that an adversary with arbitrary background knowledge cannot infer private input from the noisy results. This study proposes FinPrivacy, a privacy-preserving mechanism for fingerprint identification. This mechanism utilizes the low-rank matrix approximation to reduce the dimensionality of fingerprint and the exponential mechanism to carefully determine the value of the optimal rank. Thereafter, FinPrivacy injects Laplace noise to the singular values of the approximated singular matrix, thereby trading off between privacy and utility. Analytic proofs and results of the comparative experiments demonstrate that FinPrivacy can simultaneously enforce ɛ-differential privacy and maintain an efficient fingerprint recognition.

Today, users of social platforms upload a large number of photos. These photos contain personal private information, including user identity information, which is easily gleaned by intelligent detection algorithms. To thwart this, in this work, we propose an intelligent algorithm to prevent deep neural network (DNN) detectors from detecting private information, especially human faces, while minimizing the impact on the visual quality of the image. More specifically, we design an image privacy protection algorithm by training and generating a corresponding adversarial sample for each image to defend DNN detectors. In addition, we propose an improved model based on the previous model by training an adversarial perturbation generative network to generate perturbation instead of training for each image. We evaluate and compare our proposed algorithm with other methods on wider face dataset and others by three indicators: Mean average precision, Averaged distortion, and Time spent. The results show that our method significantly interferes with DNN detectors while causing weak impact to the visual quality of images, and our improved model does speed up the generation of adversarial perturbations.

The analysis for social networks, such as the sensor-networks in socially networked industries, has shown a deep influence of intelligent information processing technology on industrial systems. The large amounts of data on these networks raise the urgent demands of analyzing the topological content effectively and efficiently in Industrial Internet of Things. One of the ways to locate important information amongst such large troves of data is to recognize influential nodes. In this article, we examine an intelligent way to recognize the influence of such nodes automatically. Motivated by the concepts of system controllability and observability from control theory, we introduce a novel method to evaluate nodes from two different aspects, namely, the ability of “observe” information on the network (i.e., observability), and the ability to propagate information to other nodes (i.e., controllability). We propose a unified data mining framework that incorporates content analysis with nodes behavioral tendencies, and show that it is able to outperform competitive baselines in recognizing influential nodes in networks. We also show that it is important to detect the presence of spammer nodes within networks, which might otherwise be wrongly recognized as influential nodes. The experimental results demonstrate the superiority of the proposed approach in comparison with baseline methods.

There are great challenges in performing graph coloring on GPU in general. First, the long-tail problem exists in the recursion algorithm because the conflict (i.e., different threads assign the adjacent nodes to the same color) becomes more likely to occur as the number of iterations increases. Second, it is hard to parallelize the sequential spread algorithm because the color allocation depends on the adjoining iteration. Third, the atomic operation is widely used on GPU to maintain the color list, which can greatly reduce the efficiency of GPU threads. In this article, we propose a two-stage high-performance graph coloring algorithm, called Feluca , aiming to address the above challenges. Feluca combines the recursion-based method with the sequential spread-based method. In the first stage, Feluca uses a recursive routine to color a majority of vertices in the graph. Then, it switches to the sequential spread method to color the remaining vertices in order to avoid the conflicts of the recursive algorithm. Moreover, the following techniques are proposed to further improve the graph coloring performance. i) A new method is proposed to eliminate the cycles in the graph; ii) a top-down scheme is developed to avoid the atomic operation originally required for color selection; and iii) a novel color-centric coloring paradigm is designed to improve the degree of parallelism for the sequential spread part. All these newly developed techniques, together with further GPU-specific optimizations such as coalesced memory access, comprise an efficient parallel graph coloring solution in Feluca. We have conducted extensive experiments on NVIDIA GPU. The results show that Feluca can achieve 1.19 – 8.39× speedup over the state-of-the-art algorithms.

Internet of Things and the related computing paradigms, such as cloud computing and fog computing, provide solutions for various applications and services with massive and high-dimensional data, while producing threats to the personal privacy. Differential privacy is a promising privacy-preserving definition for various applications and is enforced by injecting random noise into each query result such that the adversary with arbitrary background knowledge cannot infer sensitive input from the noisy results. Nevertheless, existing differentially private mechanisms have poor utility and high-computation complexity on high-dimensional data because the necessary noise in queries is proportional to the size of the data domain, which is exponential to the dimensionality. To address these issues, we develop a compressed sensing mechanism (CSM) that enforces differential privacy on the basis of the compressed sensing (CS) framework while providing accurate results to linear queries. We derive the utility guarantee of CSM theoretically. An extensive experimental evaluation on real-world data sets over multiple fields demonstrates that our proposed mechanism consistently outperforms several state-of-the-art mechanisms under differential privacy.

The early diagnosis of pulmonary cancer can significantly improve the survival rate of patients, where pulmonary nodules detection in computed tomography images plays an important role. In this article, we propose a novel pulmonary nodule detection system based on convolutional neural networks (CNN). Our system consists of two stages, pulmonary nodule candidate detection and false positive reduction. For candidate detection, we introduce Iterative Self-Organizing Data Analysis Techniques Algorithm (ISODATA) to Faster Region-based Convolutional Neural Network (Faster R-CNN) model. For false positive reduction, a three-dimensional convolutional neural network (3D-CNN) is employed to completely utilize the three-dimensional nature of CT images. In this network, Focal Loss is used to solve the class imbalance problem in this task. Experiments were conducted on LUNA16 dataset. The results show the preferable performance of the proposed system and the effectiveness of using ISODATA and Focal loss in pulmonary nodule detection is proved.

The wireless video transmission environment of unmanned aerial vehicles (UAVs) is complex and unstable given the high mobility and changeable working conditions of UAVs, which lead to burst and consecutive errors and high error rates. A compressed video stream is extremely sensitive to transmission errors, such that even a single bit error sharply degrades the video quality. Hence, we propose an intraframe pixel-row-interleaved error concealment algorithm that interleaves pixel rows to generate high similarity in different parts of a frame, thereby achieving intraframe error resilience. Subsequently, we suggest an interframe time-field-interleaved alternative motion-compensated prediction that allows for automatic error elimination and recovers at least four consecutive frames in wireless video communications. The experiments demonstrate that the proposed algorithms recover frames with excellent subjective and objective effects. Moreover, these algorithms can provide reliable and robust video transmission for UAVs.

Location-based services are raising remarkable convenience to our daily life while seriously threatening the location privacy of individuals. Differential privacy provides a promising privacy definition for location data. It is enforced by injecting random noise into each location such that the level of privacy and utility provided by this sanitization when querying an LBS is quantified and controlled. However, the primitive differential privacy overlooks data errors, which constantly exist in real-life location data, thereby potentially deviating a specified indistinguishability. Therefore, we determine the impact of data errors on the indistinguishability to address the abovementioned issue. Then, we design an equivalent mechanism to enforce differential privacy and analyze its privacy and utility. Extensive experimental evaluation on real-world datasets demonstrates that our proposed equivalent mechanism consistently outperforms several state-of-the-art mechanisms in data utility at the same privacy level.

The advancement of data communication technologies promotes widespread data collection and transmission in various application domains, thereby expanding big data significantly. Sensitive information about individuals, which is typically evident or hidden in data, is prone to various privacy attacks and serious risks of privacy disclosure. Corresponding approaches to data privacy preservation have been proposed to provide mechanisms for preserving data privacy while pubilishing useful information or mining valuable information from sanitized data. In this work, we present a comprehensive survey of privacy preservation in big data from the communication perspective. Specifically, we cover the fundamental privacy-preserving framework and privacy-preserving technologies, particularly differential privacy. We also survey the adaptations and variants of differential privacy for different emerging applications and the challenges to differential privacy. In addition, we provide future research directions about privacy preservation in communication field.

The combination of artificial intelligence methods and IoT based sensor data will play a critical and crucial role in various environments. Flight landing safety is a research hotspot of aviation field for a long time. Accurately predicting the landing speed is conducive to reducing the landing accidents. In this paper, we proposed an accurate aircraft landing speed prediction model based on the long-short term memory (LSTM) with flight sensor data. Firstly, we analyze and pre-process the dataset with statistical method including randomness tests and stationary tests. Secondly, we design the features by random forest algorithm and reduce the dimensionality of features with principal component analysis. Thirdly, we develop a deep architecture based on long-short term memory to predict the aircraft landing speed. Experiment results prove that it has better performance with higher prediction accuracy compared with the state of the art, indicating that the proposed model is accurate and effective. The findings are expected to be applied into flight operation practice for further preventing of landing accidents and improving the air management for air traffic controllers.

Achievement of good reconstruction performance by most of existing greedy algorithms is possible only when signal sparsity has been known well in advance. However, it is difficult in practice to ensure signal sparsity making the reconstruction performance of the greedy algorithms stable. Moreover, some greedy algorithms with previous unknown signal sparsity are time-consuming in the process of adaptive adjustment of signal sparsity, and thereby making the reconstruction time too long. To address these concerns, the greedy algorithm from signal sparsity estimation proposed in this paper. Based on the restricted isometry property criterion, signal sparsity is estimated before atoms selection and the step size of atoms selection adjusted adaptively based on the relations between of the signal residuals in each iteration. The research which solves the problem of sparsity estimation in the greedy algorithm provides the compressed sensing available to the applications where the signal sparsity is un-known. It has important academic and practical values. Experimental results demonstrate the superiority of the performance of proposed algorithm to the greedy algorithms with previous unknown signal sparsity, no matter on the performance stability and reconstruction precision.

Particle Swarm Optimization (PSO) algorithms often face premature convergence problem, specially in multimodal problems as it may get stuck in specific point. In this paper, we have enhanced Dynamic-PSO i.e. and an extention of our earlier research work. This newly proposed algorithm Guided Dynamic-PSO (GDPSO) also targets the particles whose personal best get stuck i.e. their personal best does not improve for fixed number of iterations similar to DPSO, however a new approach is proposed for replacing personal bests of such particles. The replacement of this new personal best is done on the basis of sharing fitness so that better diversity can be provided to avoid the problem. The performance of GDPSO has been compared with PSO and its variants including DPSO over 24 benchmark functions provided by Black-Box Optimization Benchmarking (BBOB 2015). Results show that the performance of GDPSO is better in comparison with other peer algorithms. Further effectiveness of GDPSO is demonstrated in digital image watermarking. Digital image watermarking schemes primarily focus on providing good tradeoff between imperceptibility and robustness along with reliability in watermarked images produced for wide variety of applications. To support watermarking scheme in achieving this tradeoff, suitable watermark strength is identified in the form of scaling factor using GDPSO for colored images. Results achieved through GDPSO are compared with PSO and other widely accepted variants of PSO over different combination of host and watermark images. Experiment results demonstrate that performance of underline watermarking scheme when used with GDPSO, in terms of imperceptibility and robustness, is better than other variants of PSO.

As data-intensive cluster computing systems like MapReduce grow in popularity, there is a strong need to promote the efficiency. Recurring queries, repeatedly being executed for long periods of time on rapidly evolving data-intensive workloads, have become a bedrock component in big data analytic applications. Consequently, this paper presents optimization strategies for recurring queries for big data analysis. Firstly, it analyzes the impact of recurring queries efficiency by MapReduce recurring queries model. Secondly, it proposes the MapReduce consistent window slice algorithm, which can not only create more opportunities for reuse of recurring queries, but also greatly reduce redundant data while loading input data by the fine-grained scheduling. Thirdly, in terms of data scheduling, it designs the MapReduce late scheduling strategy that improve data processing and optimize computation resource scheduling in MapReduce cluster. Finally, it constructs the efficient data reuse execution plans by MapReduce recurring queries reuse strategy. The experimental results on a variety of workloads show that the algorithms outperform the state-of-the-art approaches.

Measurement matrix and signal reconstruction algorithm are the key factors influencing the performance of signal reconstruction. However, the reconstruction performance of the existing matching pursuit algorithms, the most popular reconstruction algorithms, is closely related to the signal sparsity, which is hard to determinate apriori. As well, the researches on the reconstruction algorithms are developed independently with the design of measurement matrix. So, in this paper, we originally conduct a joint study of design of measurement matrix and signal reconstruction algorithm. RIP criterion is used to quantitatively analyze the relationship between the signal sparsity and the measurement matrix, and then an efficient joint compression and sparsity estimation matching pursuit (JCSEMP) algorithm is proposed. JCSEMP algorithm constructs a chaotic measurement matrix, a sparsity estimation algorithm based on the chaotic measurement matrix, and a variable atom selection criterion which use the variation between the residuals to adaptively adjust the number of atoms to select. Experimental results demonstrate that this algorithm can provide a better reconstruction performance and a lower reconstruction period.

A key challenge in computer networking is how to organize network topology effectively among a large number of servers in the cloud storage system. In a cloud environment, the topology, which is different from the underlying topology, may be established in any form at any potential edge peers. The cloud content delivery network (CDN) always faces problems of complex distributed path creation, cache update, load balancing, etc. To address the problem as a static content delivery, we propose an Improved Heuristic Genetic Algorithm for Static Content Delivery in Cloud Storage (IHGA-SCDCS) based on a resource management model and cost model. The static content delivery in cloud storage is abstracted into mathematical model for set solving problem, which is then solved by an improved Genetic Algorithm (GA). Finally, the optimal solution is reduced to an optimal content delivery program. The simulation experiment, based on CloudSim, shows that IHGA-SCDCS can effectively obtain optimal solution while reducing delivery cost.

The challenge for real-life traffic sign detection lies in recognizing small targets in a large and complex background, making state-of-the-art general object detection methods not work well in both detection speed and precision. The existing deep learning models for traffic signs detection fail to use the fixed feature of the targets. This paper proposes a novel end-to-end deep network that extracts region proposals by a two-stages adjusting strategy. Firstly, we introduce an AN (Attention Network) to Faster-RCNN for finding all potential RoIs (Regions of Interest) and roughly classifying them into three categories according to colour feature of the traffic signs. Then the FRPN (Fine Region Proposal Network) produces the final region proposals from a set of anchors per feature map location extracted by the AN. We also modify the model by (1) adding a deconvolutional structure to convolutional layers to fit the small size of targets, and (2) replacing the classifier with three softmax corresponding to three coarse categories obtained by the AN. Our method is evaluated on two publicly available traffic sign benchmarks which are collected in real road condition. The experiments show our method generates only 1/14 of the anchors generated by Faster-R-CNN so the detection speed is increased by about 2 fps with ZF-Net and it reaches an average mAP of 80.31% and 94.95% in two benchmarks, 9.69% and 7.88% higher than Faster-R-CNN with VGG16, respectively.

Performance diagnosing for HPC applications can be extremely difficult due to their complicated performance behaviors. One hand, developers used to identify the potential performance bottlenecks by conducting detailed instrumentation, which may introduce significant performance overheads or even performance deviations. On the other hand, developers can only conduct small numbers of application runs for profiling the performance with the limitations on both computing resources and time duration. Meanwhile, the performance bottlenecks of HPC applications may vary with the degree of parallelism. To address these challenges, our paper proposes a systematic performance diagnosing method focusing on building an accurate and interpretable performance model with performance counters. Our method is able to diagnose the HPC application scaling issues by predicting its runtime and performance behaviors in different functions. After applying this modeling method on three real-world HPC applications, HOMME, CICE and OpenFoam, our evaluations show that our diagnosing method based on the performance model has the ability to diagnose the potential scaling issues, which is typically missed by the traditional performance diagnosing method and achieves about 10% prediction errors in a scale of 4096 MPI ranks on two problem sizes.

Since software aging has been proposed for decades, resource consumption parameters and performance parameters have been used to identify whether running a commercial web server has been in aging state or failure state. However, the relationship between workload parameters and resource consumption parameters has not been analyzed and also sensitivity between resource consumption parameters and workload parameters has not been studied before. In this work, we give an experimental case study about resource consumption parameters and workload parameters in an Internet Information Services. Firstly, we use fitted resource consumption parameter to learn the relationship between workload parameters and resource consumption parameters through visual observation and calculation. Secondly, sensitivity analysis is used to find how resource consumption parameter changes when deleting one workload parameter at a time. Thirdly, the regression tree based on a risk estimate is used to forecast resource consumption. In the experiments, we see that almost all the parameters present nonlinear feature through visual observation. And we find that some workload parameters are redundant for fitting resource parameters by using sensitivity analysis. Our proposed regression tree is better than artificial neural network by using mean absolute error.

A weighted fusion method of D-S evidence theory in decision making is proposed to aim at the problem of lacking in the distribution of trust, data processing and precision in D-S evidential theory. The method of gesture recognition based on Kinect and sEMG signal are established. Weighted D-S evidence theory is used to fuse Kinect and sEMG signals and the simulation experiment is made respectively. The stimulation results show that comparing with other experimental methods, the decision fusion method based on weighted D-S evidence theory has higher utilization efficiency and recognition rate.

In the big data era, much real-world data can be naturally represented as graphs. Consequently, many application domains can be modeled as graph processing. Graph processing, especially the processing of the large-scale graphs with the number of vertices and edges in the order of billions or even hundreds of billions, has attracted much attention in both industry and academia. It still remains a great challenge to process such large-scale graphs. Researchers have been seeking for new possible solutions. Because of the massive degree of parallelism and the high memory access bandwidth in GPU, utilizing GPU to accelerate graph processing proves to be a promising solution. This article surveys the key issues of graph processing on GPUs, including data layout, memory access pattern, workload mapping, and specific GPU programming. In this article, we summarize the state-of-the-art research on GPU-based graph processing, analyze the existing challenges in detail, and explore the research opportunities for the future.

The flying ad hoc network (FANET) is a new paradigm of wireless communication that governs the autonomous movement of UAVs and supports UAV-to-UAV communication. A FANET can provide an effective real-time communication solution for the multiple UAV systems considering each flying UAV as a router. However, existing mobile ad hoc protocols cannot meet the needs of FANETs due to high-speed mobility and frequent topology change. In addition, the complicated flight environment and varied flight tasks lead to the traditional built-in-rules protocols no longer meeting the demands of autonomy. Hence, we have proposed adaptive hybrid communication protocols including a novel position-prediction-based directional MAC protocol (PPMAC) and a self-learning routing protocol based on reinforcement learning (RLSRP). The performance results show that the proposed PPMAC overcomes the directional deafness problem with directional antennas, and RLSRP provides an automatically evolving and more effective routing scheme. Our proposed hybrid adaptive communication protocols have the potential to provide an intelligent and highly autonomous communication solution for FANETs, and indicate the main research orientation of FANET protocols.

The access security of wireless devices is a serious challenge in present wireless network security. Radio frequency (RF) fingerprint recognition technology as an important non-password authentication technology attracts more and more attention, because of its full use of radio frequency characteristics that cannot be imitated to achieve certification. In this paper, a RF fingerprint identification method based on dimensional reduction and machine learning is proposed as a component of intrusion detection for resolving authentication security issues. We compare three kinds of dimensional reduction methods, which are the traditional PCA, RPCA and KPCA. And we take random forests, support vector machine, artificial neural network and grey correlation analysis into consideration to make decisions on the dimensional reduction data. Finally, we obtain the recognition system with the best performance. Using a combination of RPCA and random forests, we achieve 90% classification accuracy is achieved at SNR $$\ge $$ 10 dB when reduced dimension is 76. The proposed method improves wireless device authentication and improves security protection due to the introduction of RF fingerprinting.

Detection plays an important role in improving the performance of multi-object tracking (MOT), but most recently MOT works mainly focus on association algorithm and usually ignore the detections. To assist in associating object detections and to overcome detection failures, in this paper, we explore the low-rank-based foreground detection method to refine the detections and show it can significantly lead a better tracking result in online multi-object tracking. Firstly, the low-level pixel information from low-rank foreground segmentation and high-level detection responses from object detector are combined to form an overcomplete detections set, which serves as input for the tracking-by-detection-based multi-object tracking. Then, the predicted object location in online tracking as a prior to feedback for the foreground segmentation in sparse approximation for future frames can improve the foreground detection performance. Finally, to effectively solve the data association problem in online MOT, two-step data association relies on tracklet confidence is used to associate the detections and generate long trajectories since the existing trajectories provide a reliable history to support their presence in current frame. The experimental results in public pedestrian tracking datasets show that our detection optimization strategy can help to improve the tracking performance compared with several state-of-the-art multi-object trackers, with improved recall, precision, FP, FN and MOTA, MOTP results.

In the today’s scenario, the number of online song repositories such as iTunes, Hungama.com , etc. is increasing day-by-day. The reason for this can be attributed to the exponential growth in the Internet users in the past few years. These song repositories store huge number of songs (mostly in millions) and charge their users for listening and downloading them. With increased number of users requires more enhanced security measures to protect such vulnerable songs repository. Any breach in security of such song repositories would not only cause huge financial loss but also copyright infringement for the owners. Therefore, in this paper we have presented a novel and efficient approach for providing security and privacy to huge and vulnerable songs repository using visual cryptography. Presented approach not only provides confidentiality to the songs but also provides integrity verification with access control to the songs repository. We have also removed various basic security constraints of (2, 2) visual cryptography existed in most of the state of art approaches like meaningless pattern of the shares, explicit codebook requirement, contrast loss, lossy recovery etc which are eliminated in the proposed approach.

Large-scale tridiagonal matrix solvers based on heterogeneous systems currently cannot balance computational efficiency and numerical stability when solving a non-diagonally dominant matrix. A tridiagonal solver combined central processing unit with graphics processing unit is proposed, based on SPIKE2 as a solver framework, a simplified SPIKE algorithm as a central processing unit component, and a diagonal pivot algorithm as a graphics processing unit component. The solver performance is further improved by using a data-layout-transformation mechanism to obtain continuous addresses, reducing memory communication using constant memory to store unchanged data in the calculation process, and employing a kernel-fusion mechanism to reduce power consumption of graphics processing unit. For a diagonally dominant matrix, extended Thomas algorithms and cycle reduction to replace the graphics processing unit component are proposed in the solver. Experimental results show that the tridiagonal matrix solver in this paper can effectively consider both numerical stability and computational efficiency, and reduce total power consumption while improving memory efficiency.

IEEE Pulmonary cancer is one of the most dangerous cancers with a high incidence and mortality. An early accurate diagnosis and treatment of pulmonary cancer can observably increase the survival rates, where computer-aided diagnosis systems can largely improve the efficiency of radiologists. In this paper, we propose a deep automated lung nodule diagnosis system based on three-dimensional convolutional neural network (3D-CNN) and support vector machine (SVM) with multiple kernel learning (MKL) algorithms. The system not only explores the computed tomography (CT) scans, but also the clinical information of patients like age, smoking history and cancer history. To extract deeper image features, a 34-layers 3D Residual Network (3D-ResNet) is employed. Heterogeneous features including the extracted image features and the clinical data are learned with MKL. The experimental results prove the effectiveness of the proposed image feature extractor and the combination of heterogeneous features in the task of lung nodule diagnosis.

Multimedia networks hold the promise of facilitating large-scale, real-time data processing in complex environments. Their foreseeable applications will help protect and monitor military, environmental, safety-critical, or domestic infrastructures and resources. Cloud infrastructures promise to provide high performance and cost effective solutions to large scale data processing problems. This paper focused on the outlier detection over distributed data stream in real time, proposed kernel density estimation (KDE) based outlier detection algorithm KDEDisStrOut in Storm, firstly formalized the problem of outlier detection using the kernel density estimation technique and update the transported data incrementally between the child node and the coordinator node which reduces the communication cost. Then the paper adopted the exponential decay policy to keep pace with the transient and evolving natures of stream data and changed the weight of different data in the sliding window adaptively made the data analysis more reasonable. Theoretical analysis and experiments on Storm with synthetic and real data show that the KDEDisStrOut algorithm is efficient and effective compared with existing outlier detection algorithms, and more suitable for data streams.

Based on the traditional segmentation algorithms, this paper proposes unsupervised video segmentation approach. The proposed algorithm applies superpixel to indicate the movement foreground and uses the static features of current frame and the relevant features of adjacent frames to compute the weight. It also brings in the mechanism of superpixel color features match restriction and motion relevance match restriction. The experiment result shows this algorithm can achieve the segmentation of video pictures and effectively solve the problem of over-segmentation.

Skeletal bone age assessment is a widely used standard procedure in both disease detection and growth prediction for children in endocrinology. Conventional manual assessment methods mainly rely on personal experience in observing X-ray images of left hand and wrist to calculate bone age, which show some intrinsic limitations from low efficiency to unstable accuracy. To address these problems, some automated methods based on image processing or machine learning have been proposed, while their performances are not satisfying enough yet in assessment accuracy. Motivated by the remarkable success of deep learning (DL) techniques in the fields of image classification and speech recognition, we develop a deep automated skeletal bone age assessment model based on convolutional neural networks (CNNs) and support vector regression (SVR) using multiple kernel learning (MKL) algorithm to process heterogeneous features in this paper. This deep framework has been constructed, not only exploring the X-ray images of hand and twist but also some other heterogeneous information like race and gender. The experiment results prove its better performance with higher bone age assessment accuracy on two different data sets compared with the state of the art, indicating that the fused heterogeneous features provide a better description of the degree of bones' maturation.

In general, dynamic traffic signal control or smooth driving benefits energy saving and CO2 emissions. However, most of studies in the past lack the integrated optimization of traffic signals control and instantaneous vehicle motion states simultaneously on energy saving and CO2 emission reduction. In this paper, a bi-level optimization model is proposed to minimize fuel consumption and CO2 emissions by considering real-time traffic signal control in road side unit (RSU) and instantaneous vehicle motion states optimization in on-board unit (OBU) synthetically. The RSU communicates with OBUs by vehicle wireless communication networks. Then, the traffic signal scheme in RSU is optimized with the received vehicle motion data sent from OBUs, and the system in vehicles optimize their speed and acceleration with the received traffic signal scheme sent from RSU. The simulation results indicate that the proposed model outperforms the existing Maximize Throughput Model (MaxTM) up to 10% in reducing fuel consumption and CO2 emissions especially when the traffic is heavy.

Axial and visibility graph analysis models are combined with GIS and Space Syntax to study the corresponding relationship between street network and specific city life in Hankou, China, based on three scales - city, district, and community, so as to interpret the hidden structure out of complex urban form through the spatial logic of street network. In this paper, a quantitative analysis is made on selected parameters in Space Syntax, including Integration, Choice, density of road, and Ht index. The result indicates that there is certain correlation among these parameters. Moreover, these parameters also present certain changing patterns along with the increase of analysis radius. The results reveal that the street network of Hankou presents a multi-hierarchical structure and spatial wholeness from all these three scales. This characteristic creates vitality and diversity while maintaining the feature of wholeness for the urban space. Research has turned out that community-scale street network performs a positive role to keep neighborhoods alive. Therefore, it is an important strategy for maintaining the vitality of urban space and realizing coordination and unification between parts and the whole. It is proposed in this paper that a good street network is a key factor for carrying forward urban context of Hankou. Undoubtedly, it is worthy of reflecting on large-scale demolition of existing urban space, especially in historical areas. Afterwards, the paper proposes a hierarchically synergetic planning strategy based on the analysis. More attention should be paid to street network to preserve diversity, continuity, and integrity, and finally archive holistically sustainable development within a city.

K-Means is the most popular clustering algorithm with the convergence to one of numerous local minima, which results in much sensitivity to initial representatives and noise point because of its distance based judgment. Density-based clustering algorithm is not sensitive to outliers and noise data, but it's difficult to present the high dimensional data as well as the changes of the data density. Grid-based method is fast, but may reduce the quality and accuracy of the cluster. However, this paper proposes a novel density based initial center optimization algorithm (DBICO) to choose the initial center, which by means of the local optimality and sensitivity of density-based clustering a lgorithm and grid-based method. The core idea is to divide the dataset into several cubes and merge some cubes, delete the noise points according to the density, calculate the initial center point and then clustering the dataset. Doing this can reduce the number of iterations, and avoid the disadvantages of the K- Means algorithm results differ due to the different initial points. Theoretic analysis and experimental demonstrations show that the algorithms this paper proposed outperforms existing algorithms in clustering quality, and it was proved fruitful applications in the logistics.

In order to save the computer data effectively, we should not only check the integrity for the data integrity, but also check storage system to recover data in a timely manner to reduce losses to a minimum, to prevent the recover fails when the fault occurred. The viability and integrity of the system to recover the data is determined by the recovery algorithm, so the key to recover the data is how to design the recovery algorithm. This paper presented an advanced recovery techniques based on the traditional recover algorithm which based on the log, then the specific data structure and algorithm implementation is given at the last of this paper.

Intrusion Detection is an important component of the security of the Network, through the critical information of the network and host system, it can determine the user’s invasion of illegal and legal acts of the user misuse of resources, and make an adequate response. According to the problem, which machine learning anomaly detection effect is not ideal when the user behavior changes and a separate anomaly detection. Based on the Bayesian inference anomaly detection, applying the Bayesian inference of statistical methods to machine learning anomaly detection, this paper established a decision tree corresponding to the method. This method overcomes the satisfactory of the anomaly detection individual test, and improves the machine learning in the predictive ability of anomaly detection and anomaly detection efficiency.

Rational location of logistics distribution center can reduce costs of transportation, operation and construction. In the research on location of practical logistics distribution centers, the accurate algorithms can not used for solution in general, so the heuristic methods such as genetic algorithm should be adapted. Based on construction of logistics distribution center location model, the paper improved on simple genetic algorithm with best conserving and no-replacing residue stochastic choose strategies. The solution to material location example of logistics distribution center was also provided. The example analysis indicates that it is more effective to improve the selection operator with the combination of best conserving and no-replacing residue stochastic choose strategies.