Abstract:Map matching is a key preprocessing step in the location-based service to match GPS points into a digital road network. Data analysis on the map matched trajectory data can be used to facilitate many real city computing applications such as intelligent transportation system and trip planning. This survey provides a systematic summary of existing research achievements of map matching. With the rapid development of urban traffic, the cost of acquiring and processing vehicle location information is increasing, low-sampling-rate GPS tracking data is growing, and the accuracy of existing algorithms is not adequate. In recent years, map matching algorithm based on hidden Markov model (HMM) has been widely studies. HMM can smoothly assimilate noisy data with path constraints by choosing a maximum likelihood path. The accuracy of HMM-based algorithms can reach 90% under certain conditions, which confirms the validity of map matching algorithm based on HMM at low sampling rate. A perspective of future work in this research area is also discussed.

Abstract:Deep neural networks have continually surpassed traditional methods on a variety of computer vision tasks. Though deep neural networks are very powerful, the large number of weights consumes considerable storage and calculation time, making it hard to deploy on resource-constrained hardware platforms such as mobile system. The number of weights in deep neural networks represents the complexity to an extent, but not all the weights contribute to the performance according to recent researches. Specifically, some weights are redundant and even decrease the performance. This survey offers a systematic summarization of existing research achievements of the domestic and foreign researchers in recent years in the aspects of network pruning, network distillation, and network decomposition. Furthermore, comparisons of compression performance are provided on several public deep neural networks. Finally, a perspective of future work and challenges in this research area are discussed.

Abstract:Policy gradient methods have been extensively studied as a solution to the continuous space control problem. However, due to the presence of high variance in the gradient estimation, policy gradient based methods are restricted by low sample data utilization and slow convergence. Aiming at solving this problem, utilizing the framework of actor-critic algorithm, a true online incremental natural actor-critic (TOINAC) algorithm, which takes advantage of the natural gradient that is superior to conventional gradient, and is based on true online time difference (TOTD), is proposed. In the critic part of TOINAC algorithm, the efficiency of TOTD is adopted to estimate the value function, and in the actor part of TOINAC algorithm, a novel forward view is introduced to compute and estimate natural gradient. Then, eligibility traces are utilized to turn natural gradient into online estimation, thereby improving the accuracy of natural gradient and efficiency of the method. The TOINAC algorithm is used to integrate with the kernel method and normal distribution policy to tackle the continuous space problem. The simulation tests on cart pole, Mountain Car and Acrobot, which are classical benchmark tests for continuous space problem, verify the effeteness of the algorithm.

Abstract:Layout design of satellite module is not only a complex coupling system design problem but also a special optimization problem. It is considered to be NP-hard. The most challenge of solving this problem is that the objective function to be optimized is characterized by a multitude of local minima separated by high-energy barriers. The Wang-Landau (WL) sampling method is an improved Monte Carlo method, which has been successfully applied to solve the protein structure prediction and other optimization problems. Taking satellite layout design as case study, this paper introduces the WL sampling method to solve the rectangular packing problem. In order to guide the WL sampling algorithm to random walk effectively in solution space, rectangular objects-oriented heuristic layout update strategies are proposed. To accelerate the search for the global optimal layout, the gradient method is executed for local search once the Monte-Carlo sweep produces a new layout. By incorporating the local search mechanism and heuristic layout update strategies into the WL sampling algorithm, a heuristic Wang-Landau sampling algorithm is constructed to solve the arbitrary rectangular packing problem with the static non-equilibrium constraint. By adding a static non-equilibrium penalty term on the basis of the extrusive elastic energy, and adopting the translation of the center of mass, the static non-equilibrium constraints of the whole system can be satisfied. Furthermore, to improve the efficiency of the algorithm significantly, an improved finite-circle method is presented to judge and calculate the overlapping depth among objects. The computational results of two sets of benchmarks consisting of ten representative instances from the literature show that the proposed packing algorithm is effective.

Abstract:As service-oriented technology and cloud computing technology continue to mature, and especially the service-oriented architecture (SOA) continue to improve and proliferate, content Web services have become widely used. In order to take full advantage of general Web services and overcome the problem of limited functionality of individual Web service, the industry embraces approach of combining multiple atomic Web services in accordance with certain rules and business logic to provide more functionality and more powerful services, enabling the increment and reuse of Web services. To ensure the quality of the Web service composition, comprehensive and adequate testing of the Web service composition is required. However, due to the dynamic and distributed features of Web service composition, the testing techniques and methods are different from those of traditional software, thus bringing many challenges. This paper systematically summarizes and analyzes test case generation techniques, regression testing techniques, test execution and measurement methods in Web service composition testing research in recent years. The paper also provides an analysis and outlook on the issues that need to be studied in the Web service composition testing.

Abstract:Context-Aware applications become more and more appealing. Eliciting and modeling the context-aware requirements face significant challenges due to the complexity and uncertainty of context. A variety of studies have proposed methods to tackle these challenges. The methodology of systematic literature review is used to analyze which dimensions of context have been used for better requirements elicitation and modeling in context-aware situations, as well as which methods have been used in context aware requirements elicitation and modeling to evaluating the technology transfer maturity of the methods. Future research directions in context aware requirement elicitation and modeling are also discussed.

Abstract:With the increasing of social network, social recommendation becomes hot research topic in recommendation systems. Matrix factorization based (MF-based) recommendation model gradually becomes the key component of social recommendation due to its high expansibility and flexibility. Thus, this paper focuses on MF-based social recommendation methods. Firstly, it reviews the existing social recommendation models according to the model construction strategies. Next, it conducts a series of experiments on real-world datasets to demonstrate the performance of different social recommendation methods from three perspectives including whole-users, cold start-users, and long-tail items. Finally, the paper analyzes the problems of MF-based social recommendation model, and discusses the possible future research directions and development trends in this research area.

Abstract:Pattern matching with gap constraints is one of the key issues of sequential pattern mining. Recently, most research work focuses on pattern matching with non-negative gaps, but the rule strictly limits the order that each character appears in the sequence. In order to increase the flexibility of matching while taking into account that it is more reasonable to use one-off condition in sequential pattern mining, this paper studies the pattern matching problem under general gap and one-off condition, which is NP-hard. To tackle this issue, an algorithm, named MSAING, is proposed. Firstly, the algorithm processes the pattern and sequence using the Reverse strategy to get the maximum number of matching results. Secondly, it significantly reduces the time and space overhead with linear table structure in the matching process, and improves the matching rate using the backtracking method. Finally, to further improve the efficiency of the algorithm, it determines whether internal repetition exists in the pattern or not, according to the inside_Checking mechanism. Completeness of the MSAING algorithm is proved in theory. Experimental results verify the accuracy of the matching results of the MSAING algorithm and its validity in terms of the time and space complexity.

Abstract:The inference techniques for probabilistic knowledge bases have recently attracted significant attentions. In most off-the-shelf existing systems, the inference is mainly implemented based on batch processing and thus not suited for online querying. This paper proposes an online inference approach based on approximate factors for probabilistic knowledge bases, so as to provide a way to reuse those inferred results to calculate the marginal probability for the query variable. In this approach, a subgraph is extracted first, taking the query variable as center; then some approximate factors are attached to simulate the influences from the variables outside the subgraph; and finally, the marginal probability of the query variable is calculated by the clique tree algorithm. Experiments show that compared with existing algorithms, the presented approach can achieve a better tradeoff between accuracy and time.

Abstract:Process mining is an active research topic in the cross field of process management and data mining. In an actual business environment, the recorded data of a process execution that may be supported by different computer systems is scattered into different event log files. It is necessary to merge the scattered data into one single event log file when applying current process mining techniques and tools for process mining. This mission is still challenging, however, because of the complex relationships between cases in two logs and the possible lack of information for the merging. In this paper, event log merging for process mining is regard as a type of search and optimization problems based on the formal definition, and a merging approach with a hybrid artificial immune algorithm is presented in order to achieve the event log merging with many to many relationship between cases in the two event logs. In the merging approach, the clonal selection principle is selected as its underlying principle, which requires the matching process to undergo iterations of clonal selection, hypermutation and receptor editing in order to get the best solution. The algorithm starts from an initial population produced with a heuristic approach. Two factors, occurrence frequency and temporal relation, are designed in the affinity function to evaluate the individuals in the population. In addition, immunological memory and simulated annealing are exploited to make the artificial immune merging jumping out from the trap of local optima. Experimental results show that the hybrid algorithm has good performance in merging logs with complex cases relationships, and the heuristic approach for initial population can speed the process of the evolution. This paper also discusses the data distribution methods in which the log merging problems can be distributed.

Abstract:Opinion mining (OM) of Internet reviews is one of the key issues in text analysis. As the rapid growth of the Internet reviews, users pay more attention to all this fine-grained information when browsing comments. Therefore, aspect-level OM can help consumers make better decisions. In last decade, researchers conducted opinion extraction and analysis on a large number of Internet reviews corpus, and have achieved fruitful research results and broaden the scope of application. There were also some scholars conducted summaries on the present situation of OM methods. To rectify the lack of specific summaries on aspect extraction and opinion expression extraction, this paper analyzes and summarizes the recent research status of aspect-level OM on Internet reviews. The paper describes the aspect-level OM, introduces the different methods of aspect extraction and opinion expression extraction, and summarizes the evaluation measures of aspect-level OM and application values. In the end, it provides an overview of the future challenges along with a synopsis on the existing techniques. This specific survey on aspect-level OM helps to evaluate the different methods and find valuable research direction.

Abstract:Authentication and confidentiality, as well as sender and receiver anonymity are essential in broadcast communication. In this paper, an anonymous hybrid signcryption scheme with multi-receiver is proposed using identity-based cryptography. The proposal does not contain receiver's identity list, and the identity of sender is included in an identity set. Thus, it not only obtains the receiver's anonymity, but also achieves the sender's anonymity. Additionally, the proof of security and the analysis of correctness demonstrate that the scheme is secure and effective. Compared with the pre-existing schemes, the proposal enjoys better performances in many perspectives, including confidentiality, unforgeability, higher anonymity of sender and receiver and public verifiability. Moreover, the presented method can be improved to develop an efficient construction of hybrid signcryption scheme with multi-message and multi-receiver, which can obtain these security properties, such as sender and receiver anonymity, public verifiability and non-repudiation. Finally, the new variant can achieve the requirement of sending multi-message in broadcast communication.

Abstract:RGB is one of the mixed multivariate signature schemes. It can resist the current known algebraic attacks against multivariate schemes. However, similar to other MPKC (multivariate public key cryptosystem) schemes, it suffers from large public key size. In order to reduce the public key size of RGB, this study extends the idea of cyclic public key to the RGB signature scheme and proposes a new scheme-CyclicRGB mixed multivariate signature system. In comparison with the original scheme, the new scheme has smaller public key size, and it is also possible to speed up the verification process. Furthermore, experiments demonstrate that it is feasible to reduce the size of public key. The public key size of the new scheme is about 40% of that for the RGB scheme. The time required for the signature verification in CyclicRGB scheme is about 60% of the time required for the RGB scheme.

Abstract:Shape descriptor is a concise and informative representation. Feature extraction is a key step in many 3D shape analysis tasks. In recent years, feature extraction technologies of non-rigid 3D shape have attracted a lot of attentions. This paper firstly introduces the evaluation criteria and the datasets which are commonly used as benchmark in non-rigid 3D shape feature extraction. Secondly, based on extensive research on the existing literatures and the latest achievements, the paper categorizes the non-rigid 3D shape descriptors into two types:Hand-Crafted shape descriptors and learning based shape descriptors. The basic ideas, advantage and disadvantage of typical algorithms belong to each category, especially the most recent feature extraction algorithms based on deep learning are analyzed, compared and summarized. Finally, some potential future work is discussed.

Abstract:Ensemble simulation is increasingly popular in scientific domain such as climate research, weather report, mathematics and physics. Ensemble simulation data sets are usually multi-valued, multi-variate, time-variant and large in scale. Thus, analyzing such data sets is challenging. Ensemble visualization helps scientists to compare ensemble members and give overall summary to the whole data sets by utilizing visual encoding and human interaction. It thus helps scientists to explore, conclude and validate their findings. This article describes analytical tasks and strategies for organizing existing works on visualization and visual analysis on ensemble simulation data sets. The analytical tasks for ensemble simulation data sets include comparing individual members and summarizing whole ensemble, whereas the analytical strategies consist of location-based method and feature-based method. This article reviews major works in ensemble visualization. It gives explanation to their visual design, interaction approaches and application scenarios, along with a discussion of recent trends and future research directions.