Abstract:Subject extraction from a text is very important for natural languag e processing. Traditional methods mainly depend on the mode of "thesaurus plus m atch". It is not fit to process Internet news because of its limited volume and slow update speed. After analyzing the news structure carefully, this paper pres ents a new practical method to extract news subjects without thesaurus, and give the main implementing procedure. Instead of large thesaurus, it uses the specia l structure of Internet news to find the repeated strings. These repeated string s could express the news subjects very well. Experimental results show that this method can extract the most important subject strings from most of Internet new s rapidly and efficiently. Moreover, this method is equally efficient to other A sian languages such as Japanese and Korean, as well as other western languages.

Abstract:How to securely distribute a company's secret key to its n authorized departments is of much importance in information management system of E-commerce. In many cases, every authorized department in a company needs a partial or sub-secret key, and each person in the corresponding authorized department has an access level to the sub-secret key. Upon the sub-secret key of a department being cooperatively reconstructed, the company's secret key can be securely and completely reconstructed from all sub-secret keys of the departments. This paper presents such a novel multi-level secret sharing scheme. Instead of traditional method, this scheme is mainly based on the complexity of semigroup structures in which each department's sub-secret key is related to a group and each person's access level in the corresponding department is characterized by an order of an element in the group. The proposed scheme can also be employed to other scenarios where multi-level secret sharing is needed.

Abstract:In this paper, the authors first provide a close form differentiate formula to analyze the buffer requirement for zero loss of TCP in networks with high bandwidth delay product. Then, A Token Based Fast-TCP algorithm is analyzed and modeled in WANs when the window size is smaller than Bandwidth-Delay Products (BDP) or in satellite networks where window size is not allowed to be larger than BDP because link errors. It proves that Token Based ACKs delay algorithm not only reduce buffer occupancy largely, but also gain the average same throughput as TCP with infinite buffer size as well. The authors also apply the analysis to multiple connection case that is not included in previous work.

Abstract:An important issue in the design of a mobile computing system is how to manage the real-time locations of mobile clients. In the existing commercial cellular mobile computing systems, a two-tier architecture is used. However, the two-tier architecture is not scalable and is not suitable to the new mobile computing applications in which there are a very large of mobile users. In the literatures, a hierarchical database structure is proposed. The location databases of different cells are organized into a tree structure to facilitate the search of mobile clients. Although this architecture can distribute the update and searching workload amongst the location databases in the system, it has the problem of heavy location update overhead and long search delay. This paper studies how to generate location update based on the distance-based update method in the tree structure. A location update generation method is proposed to calculate the optimal distance threshold with the objective to minimize the total location management cost.

Abstract:Model checking is an algorithmic technique for checking if a concurrent system satisfies a given property expressed in an appropriate temporal logic. LTLC(linear temporal logic with clocks) is a continuous-time temporal logic proposed for the specification of real-time systems. It is a real-time extension of the temporal logic LTL. In this paper, the model checking problem for LTLC is discrssed and a reduction from LTLC model checking to LTL model checking is presented. This reduction will enable us to use the existingLTL model checking tools for LTLC model chcking.Owing to the to the fact LTLC canexpress both the proerties and implementations of real-time sys,the LTLC modelchecking procedures can be used for both the prperty verication and the refinement verification for real-time systems with finite locations.

Abstract:This is a review paper about the basic model approaches and model characteristics as well as the recent works on SPA. In this paper, the difference and similarities about SPA and SPN are compared and the approach of translating SPA into the SPN semantics and exploiting SPA achievements for SPN are discussed. At the end of this paper, some research directions and open problems are presented in this area.

Abstract:The exploding of information in data warehouses makes a great challenge to data warehouse research. How to increase query efficiency across web data warehouse plays an important role in data warehouse research. After analysis of several data warehouse implementation, a kind of hierarchy architecture of the web data warehouses is proposed, and an asynchronous iteration approach for query processing on web data warehouses is presented based on the hierarchy architecture of web data warehouse. In the asynchronous iteraion approach,the pipe;ining paralle processing technique is exp¹oited.During of the queries on web data warehouses,all the nodes in difference layers of web data warehouse are executed in pipelining paralle manner so that the performance of the query processing is tremen dously inmproved.The theoretical analysis shows that the asynchronous iteration approach is very efficiently for the query processing on web data warehouses.

Abstract:In this paper, a 3D medical imaging surface reconstruction scheme was proposed, which integrates segmentation and Marching Cubes algorithm. An appropriate segmentation method was applied to the image sequence according to the feature of the original medical image, then the segmentation result (binary image data set) was used as the input of MC to generate iso-surfaces. In addition, a surface-tracking algorithm based on region growing is developed, which improve the efficiency by avoiding detecting empty space.Applying the scheme,both the speed and the display performance of the system are improved.

Abstract:Execution environments (EEs) are infrastructures on which applications and mobile code are programmed, managed and executed. A new type of EE for active network based on programmable mobile softdevices is presented in this paper. Programmable units, called softdevices in the EEs, are formed according to functions of protocol sub-trees. By means of demux (demultiplexer), passive packets and active packets can find and execute corresponding codes. In these EEs, functions of active nodes are extended flexibly,user prpgramming is facilitated by rich APIs,and protocols or appications are tested and deployed dynamically.

Abstract:From the view of the software fault-tolerance, in the distributed computing environment based on the network, for the property of the military command system and bank management system and in order to satisfy the reliability requirement of these applications, based on the group communication technology and by using the process replication technology, in this paper, a nested dynamic fault-tolerant model and a dynamic fault-tolerant algorithm are proposed. When the primary service process crashes, this algorithm can dynamiclly select one of backup process as a new primary service process and assure the state of all backup service processes is consistent.

Abstract:The terabyte-level (10¹² bytes) database systems such as digital libraries need to use tape library as an online device to store and retrieve their massive data. Since the access of a tape library is often random and the tape library has low random access performance, thus it is critical to study the random I/O scheduling strategies and algorithms in order to improve the performance of tape library. In this paper we study a cost-benefit-based scheduling algorithm, and as well as give an effective estimating method of cost-benefit weight ratio.This algorithm improves the performance of tape library system under different workloads by dynamically tuning the cost-benefit weight ratio of scheduling policies according to workloads.This algorithm particularly overcomes the problem of workload-sensitive of existing scheduling algorithms,and is significantly effrctive under heavy workload.

Abstract:This paper focuses on methods of optimizing a single classifier and combining multiple classifiers by genetic algorithms (GA). The method uses the strategies of stacking. There are two steps in classical strategies of stacking, and GA is used as the second step in the method. Experimental results show that it performs well on the task of optimization. Comparing with the single algorithm, it enhances the precision. In task of combiningoptimization, it can obtain more understandable result than constituent learners.

Abstract:The parallel query optimization for multi-join expressions is one of the key factors to improve the performance of database systems. In this paper, an approach to solve the problems of the parallel query optimization for multi-join expressions by adopting GA algorithms is proposed. To improve the execution efficiency of the query processors, the authors exploit heuristics to seek the optimum parallel scheduling execution plan for multi-join expressions. The detailed testing results and performance analysis are presented.The experiment results show that the GA algorithm with heuristic knowledge is effective for parallel query processing of multi-joins,and plays an important role in improving the performance of database systems.

Abstract:Synchronization is one of the key technologies in maintaining the consistency of replicated mobile database systems. A lot of research has been focused on transactional synchronization schemes in recent years and many models and algorithms have been proposed. However, existing transactional synchronization schemes have some limitations. In order to solve these problems, a new synchronization scheme, double-timestamp transaction-level synchronization (DTSTLS) model, is put forward in this paper. DTSTLS model uti;izes a three-tire architecture.Therefore existing database products can be used as its database server,which improves the flexibility and extensibility of the system.As an asynchronous group replication method,DTSTLS model allows allows users to access local replicas of the database and to locally submit mobile transactions when the system is disconnected.The locally committed transaction are sent to the syuchronization server for conflict reconciliation and synchronization when the system is reconnected.Mobile transactions thst have executed conflicting operations will be aborted,and only those mobile transactions that do not violate the consistency of the database system can be globally committed.In this way,the consistency of the database system is ensured.Besides,Besides,Besides,through a novel timestamp strategy,DTSTLS model cuts down the resource resource consumption and reduces the communication costs.Detailed simulation experiments are carried out,and the results show that the proposed methodology not only ensures the serializability and consistency of the mobile database system,but also reduces storage consumption and communication cost.

Abstract:The research is hierarchized by advanced network infrustructure, network services, network applications and their supporting technology. Accordingly, the content covered following fields: advanced network operation and management, network and information security, network interconnection and routing, network design and testing, Chinese network environment and information discovery, typical network applications.

Abstract:The reliable communication problem in mobile agents environment hasn抰 been satisfactorily addressed in currently available systems and research. In this paper, a novel algorithm is presented on location management and reliable communication for mobile agents. In virtue of the algorithm, the messages for remote mobile agents can be trustworthily accepted. And the exactly-once semantics for message delivery is guaranteed.

Abstract:Speech segmentation is the foundation of some applications such as speech recognition and spoken document retrieval. An improved algorithm is proposed here which include: (1) GLR variance based threshold adaptive algorithm is to improve the threshold selection approach in speaker based speech segmentation under various acoustic environments;(2) BICs Detection Ability is referred to determine when BIC is effective;(3) Besides to verify the candidate segmentation points, BIC is used to calibrate their bias caused by GLR variance.Experimental results indicate that the improved algorithm is prior to the original one.

Abstract:For many KDD (knowledge discovery in databases) applications, such as fraud detection in E-commerce, it is more interesting to find the exceptional instances or the outliers than to find the common knowledge. Most existing work in outlier detection deals with data with numerical attributes. And these methods give no explanation to the outliers after finding them. In this paper, a hypergraph-based outlier definition is presented, which considers the locality of the data and can give good explanation to the outliers,and it also gives an algorithm called HOT(hypergraph-based outlier test) to find outliers by counting three measurements,the support,belongingness and deviation of size,for each vertex in the hypergraph.This algorithm can manage both numerical attributes and categorical attributes.Analysis shows that this approach can find the outliers in high-dimensionsal space effctively.

Abstract:As the process of proliferation of the Internet progresses and multicast applications are deployed for mainstream use, the need to IP multicasting technology will become critical. Now, the research of multicast technology has become a fresh new thriving academic field. Many main results converge to such fields as the multicast routing, the flow control, the congestion control and reliable multicast. But the results referred to multicast security are very few (especially in the field of group key management).In this paper,the problems of group key management and access control for large dynamic multicast groups have been researched.A scheme based on subgroup secure controllers are proposed which solves many problems of the Iolus System and the solution of WGL, and at the same time,which simplifies the scheme of access control and meets our design aims and requirements.

Abstract:A kind of spatial domain image watermarking scheme based on direct sequence spread spectrum (DSSS) codes is proposed in this paper. On the basis of establishing the communication model of digital image watermarking, the authors generate a visual mask from the cover image to keep the fidelity of the image, and then embed the spread spectrum modulated watermark into the corresponding spatial domain of the digital image. At the same time, the error correction coding technique is used to improve the anti-jamming capability of the watermark.The betection is realized by computing the correlaon between the difference image and the original spread spectrum codes.Experimental results shoe that the scheme improves the robustness and invisibility of the image watermark,while has preferable subjective effects.

Abstract:In this paper, a kind of globally convergent continuous neural network for optimization problems is presented by designing Liapunov function skillfully, it has better function and higher performance. It is capable of solving nonlinear programming problems with the constraints of equality and inequality. The proposed neural network is an extension of Newton deepest decedent method for constraint problems, it can improve the accuracy of the solutions, and its structure is simpler than the existing networks even when it is for solving positive definite quadratic programming problems.

Abstract:In the paradigm of component-based software development, the component model is regarded as the foundation of almost all tasks such as classification, retrieval and adaptation. However, to date the semanticrepresentation of component is a critical problem in the model. In this paper, a semantics intensive component model is proposed. Through a semantics defining method in principle of domain analysis on the feature space, the three views of component, that is domain space, define space and context space, are suggested to describe the semantic strure,and the feature description logics is used to express the semmantic content.Especially,both the concept of semantisizer and formal method of the feature space are one possible solution to component reuse automation and engineering.