Abstract:The research of precise execution reachability analysis focuses on figuring out the reachability between program states. It tries to find witness inputs and the execution traces that pass through the setting-up target state of certain code location by performing necessary test and verification on executable files, source code and documentation. Precise execution reachability analysis has been applied to direct testing, bug reproduction, construction of proof of concepts of vulnerabilities, verification for result of static analysis and so on. This paper provides a survey of this area. First, the corresponding constraint solving problem of precise reachability analysis is cited. Next, existing typical methods and technical difficulties about bidirectional symbolic analysis and program induction, and some technical difficulties are discussed. Then, the applications of current precise reachability analysis are classified and summarized. Furthermore, the challenges on program analysis, program induction and constraint solving are provided. Last but not least, the possible solution and future research are suggested.

Abstract:Self-Adaptation provides a promising approach to managing the complexity of modern software systems, and in particular, to enabling systems to continuously adapt themselves to uncertainty in the environment. In existing works, model-based self-adapatation is a type of widely used methods that extend the applicability of model-driven engineering techniques to the runtime environment to facilitate self-adaptation. By leveraging software models to manage the abundant information associated with runtime phenomena, model-based self-adaptation methods are able to avoid the complexity introduced by intertwining adaptation and application behaviors and therefore simplify the development of self-adaptive systems. This paper first lays out six research questions from existing studies concerning system models, environment models, relationship among models, and model-based adaptation planning methods. It then answers the research questions by investigating relevant studies. Finally, the paper offers some suggestions for future research.

Abstract:The Internet has penetrated into all aspects of human society and has greatly promoted social progress. At the same time, various forms of cybercrimes and network theft occur frequently, bringing great harm to our society and national security. Cyber security has become a major concern to the public and the government. As a large number of Internet functionalities and applications are implemented by software, software plays a crucial role in cyber security research and practice. In fact, almost all cyberattacks were carried out by exploiting vulnerabilities in system software or application software. It is increasingly urgent to investigate the problems of software security in the new age. This paper reviews the state of the art of malware, software vulnerabilities and software security mechanism, and analyzes the new challenges and trends that the software ecosystem is currently facing.

Abstract:Crowdsourced software testing is an emerging testing method which has drawn extensive attention in both industrial and academic community. This paper systematically summarizes the academic literatures and industry practice in recent years. This article first summarizes the related literatures from the perspectives of the research topics including software testing, crowdsourcing test process, experimental subjects and scale of crowdsourcing. It also compares total of 20 widely used crowdsourced software testing commercial platforms, and discusses their task domains, subjects, open call forms and performance evaluation forms. Finally, the paper presents the future trends, issues and opportunities for crowdsourced software testing.

Abstract:Twin support vector machines have drawn extensive attention for their simple model, high training speed and good performance. The initial twin support vector machine is designed for binary classification. However, multi class classification problems are also common in practice. In recent years, researchers have devoted themselves to the study of multi class twin support vector machines. Various mulit class twin support vector machines have been proposed. The study of multi class twin support vector machines has made great progress. This paper aims to review the development of multi class twin support vector machines, classify and analyze them with the respect to the basic theories and geometric meanings. According to the structures, the paper divides the machines into the following groups:"one-versus-all" strategy based multi class twin support vector machines, "one-versus-one" strategy based multi class twin support vector machines, binary tree based multi class twin support vector machines, "one-versus-one-versus-rest" strategy based multi class twin support vector machines, and "all-versus-one" strategy based multi class twin support vector machines. Although the training processes of direct acyclic graph based multi class twin support vector machines are much similar with that of "one-versus-one" based approachs, the decision processes have their own characteristics and disadvantages, and therefore they are divided into a separate group. This paper analyzes and summarizes the ideas and theories of different multi class twin support vector machines, and presents experimental results to compare the performances. This review can make it easy for novices to understand the essential differences and help to choose the suitable multi class twin support vector machine for a practical problem.

Abstract:Machine learning problems can be viewed as optimization-centric programs, and the optimization algorithm is an important tool to solve the objective function. In the era of big data, in order to speed up the training process, it is essential to design parallel and distributed optimization algorithms by multi-core computing and distributed computing technologies. In recent years, there are a lot of research works in this field, and some algorithms have been widely applied on machine learning platforms. In this paper, five common optimization algorithms, including gradient descent algorithm, second order optimization algorithm, proximal gradient algorithm, coordinate descent algorithm and alternating direction method of multiplier, are studied. Each type of algorithm is analyzed from the view of parallel and distributed respectively, and algorithms of the same type are compared by their model type, input data characteristic, algorithm evaluation and parallel communication mode. In addition, the implementations and applications of the optimization algorithm on representative scalable machine learning platforms are analyzed. Meanwhile, all the optimization algorithms introduced in this paper are categorized by a hierarchical classification diagram, which can be used as a tool to select the appropriate optimization algorithm according to the objective function type, and also to cross explore how to apply optimization algorithms to the new objective function type. Finally, the problems of the existing optimization algorithms are discussed, and the possible solutions and the future research directions are proposed.

Abstract:The rapid growth of data and workload makes centralized database systems less and less favorable to today's applications. A distributed database system can scale out dynamically to satisfy the business development. As a result, it has gained much more attention from applications. Since the needs for distributed DB became apparent, an increasing number of products have emerged and been adopted by the Web. However, due to the complexity of distributed DB systems, their designers have to trade off among several desired properties, resulting in dramatic difference in their designs and advantages. To the best of public knowledge, no one has performed a comprehensive analysis on the design space and the tradeoff choices of modern distributed DB systems. After reviewing and understanding a significant number of real world DB products, this study believe that a distributed DB system can be generally described using three dimensions-operational consistency, transactional consistency and availability. While these dimensions are not new, their concepts are somehow blurred in the literature. This paper clarifies the three concepts in the context of database, based on which can draw a sensible landscape of the existing products and technologies. The p