Detecting Integer Overflow Vulnerabilities in Binary Executables Based on Target Filtering and Dynamic Taint Tracing

The number of identified integer overflow vulnerabilities has been increasing rapidly in recent years. In this paper, a smart software vulnerability detection technology is presented, which is used for the identifica- tion of integer overflow vulnerabilities in binary executa- bles. The proposed algorithm is combined with Target fil- tering and dynamic taint tracing （TFDTT）. Dynamic taint tracing is used to reduce the mutation space and target fil- tering function is used to filter test cases during the process of test case generation. Theory analysis indicates that the efficiency of TFDTT is higher than NonTF-DTT and ran- dom Fuzzing technology. And the experiment results in- dicate that the detection technology based upon TFDTT can identify the possible integer vulnerabilities in binary program, meanwhile, it is more efficiency than other two technologies.     

Classification of Convective and Stratiform Cells in Meteorological Radar Images Using SVM Based on a Textural Analysis

This contribution deals with the discrimination between stratiform and convective cells in meteorological radar images. This study is based on a textural analysis of the latter and their classification using a support vector machine(SVM). First, we apply different textural parameters such as energy, entropy, inertia, and local homogeneity. Through this experience, we identify the different textural features of both the stratiform and convective cells. Then, we use an SVM to find the best discriminating parameter between the two types of clouds. The main goal of this work is to better apply the Palmer and Marshall Z-R relations specific to each type of precipitation.     

NSVM Decision Tree for Multi-class Classification

A presented classification algorithm of non- linear Support vector machine （SVM） decision tree extends SVM to non-linear SVM by using kernel functions and calculates the relativity separability measure between classes after non-linear mapping. This method is introduced on the basis of the SVM decision tree. As a result, the iterative error is effectively restrained and the efficiency improved accordingly. Experimental results have shown that, compared with the original SVM decision tree algorithm, the classification rate has increased greatly and the classification time decreased apparently.     

A Novel Online Least Squares SVM Based on SMO Algorithm

To improve the training speed of online Least squares Support vector machines （LS-SVM） for large scale problems, a novel training algorithm based on Se-quential minimal optimization （SMO） is proposed in this paper. First we presented the SMO-based incremental and decremental learning algorithm, which can efficiently get new solutions based on previous training results when new samples being added or less important samples being re-moved. Then we proposed the online LS-SVM based on the incremental and decremental learning algorithm. This on-line LS-SVM no only has very high training speed and high classification accuracy but also can adaptively get sparse solutions according to objective classification problems. Finally several numerical experiments show the effectiveness of the proposed algorithm.     

Face Feature Selection and Recognition Using Separability Criterion and Binary Particle Swarm Optimization Algorithm

Discrete cosine transform （DCT） is an ef- fective method to extract proper features for face recog- nition. Discrete cosine transform can only map the re- source data to another data field instead of compress data. How to select the DCT coefficients that are most effec- tive for classification is an important problem. This paper proposes a novel method to search the best discriminant combination of DCT coefficients. A feature selection al- gorithm according to the separability criterion is used to preselect the DCT coefficients, and then follows a search algorithm based on binary particle swarm optimization and support vector machine to find an optimal combination of the DCT coefficient. The performance of the algorithm is assessed by computing the recognition rate and the num- ber of selected features on ORL database and Cropped Yale database.     

A New Weighted Least Squares Support Vector Machines and Its Sequential Minimal Optimization Algorithm

Least squares support vector machines （LSSVM） is widely used in pattern recognition and artificial intelligence domain in recent years for its efficiency in classification and regression. The solution of LSSVM is an optimization problem of a Sum squared error （SSE） cost function with only equality constraints and can be solved in a simple linear system. However, its generalization performance is sensitive to noise points and outliers that are often existent in training dataset. In order to endow robustness to LSSVM, a new method for computing weight vector of error is proposed and the substituting of weighted error vector for original error vector in LSSVM gives birth to a new weighted LSSVM. The method gets weight factor by computing distance between sample and its corresponding class center. Sequential minimal optimization （SMO） algorithm is also extended to the new method for its efficient application. Comparison experiments show superiority of the new method in terms of generalization performance, robust property and sparse approximation. Especially, the new method is much faster than the other method for large number of samples.     

支持向量机在目标分类中的应用

目标分类是一个决策过程，支持向量机是一种基于统计学习理论的机器学习方法，是目前较为理想的分类工具之一。本文介绍了支持向量机的理论背景和一些新的进展，及其在理论和实际应用中面临的难题。还对支持向量机在多元分类中的应用进行了讨论，并给出了实例。     

A Novel Battery-Aware Task Scheduling Algorithm for Multiprocessor Systems

In battery-powered portable devices, the most important design goal is maximizing the battery lifetime. This paper addresses the problem of battery- aware task scheduling on multiprocessor systems to extend the battery lifetime, and proposed a Critical-path based battery-aware task scheduling （CPbBATS） algorithm with a novel efficient flow. Experiments on hundreds of task graphs each having 10 to 100 tasks show that, compared with algorithm that simply sequences tasks with their highest voltages and utilizes the slack by the latest possible task, the CPbBATS achieved 3.3%-6.1% reduction on apparent charge lost which means longer battery lifetime, 3.6%-8.3% reduction on actual charge lost which means lower energy consumption, and average 31.99% reduction on the CPU runtime which means much lower computa- tional complexity.     

Lazy Group Key Updating Scheme Based on Dynamic Tree

In group key management systems adopting lazy updating, the overhead of key updating might be amortized. In this paper, a Dynamic-tree based Lazy group key updating scheme （DT-LGKU） is presented, which consists of four algorithms Init, Update, Derive and Extract. DT-LGKU supports lazy join events and lazy leave events with unbounded number of time intervals, and provides the forward secrecy and the backward secrecy. DT-LGKU is secure if both the Large Integer Factoring problem and RSA problem are hard. The performance evaluation shows that the worst-case complexity of the cryptographic operations used in the Extract algorithms is logarithmic in the number of time intervals, and those of Init, Update, Derive are constant.     

The On-line Calibration Technology of System Delay Based on Dynamic Delay Cancellation for Generalized TT＆C Channel Simulator

An on-line system delay calibration method based on dynamic cancellation for generalized high-precision Tracking, telemetry and command （TT＆C） channel simulator is proposed. This method manages to estimate the time-varying system delay in real time through the coupling signal of input signal and simulated output signal, and then modifies the simulated parameters by the estimated value. With this method, it effectively avoids the effect of time-varying system delay caused by temperature drift, aging of components and other factors on high precision simulation. In this paper, the dynamic cancellation technology, which is the kernel and foundation of this method, is presented to eliminate the simulated motion law between the input signal and output signal. The time delay estimation method based on cross correlation and area barycenter arithmetic is introduced to estimate the time-varying system delay. The simulation results show the validity and high-precision performance.     

An Authenticated Asymmetric Group Key Agreement for Imbalanced Mobile Networks

A Group key agreement （GKA） protocol enables a group of communicating parties to negotiate a common secret key over an open, untrusted network. The design goal of GKA is to achieve secure group communication, which is an important research issue for mobile communication. The conventional （symmetric） GKA protocol allows a group of members to establish a common secret key for imbaianced mobile networks. However, only the group members can broadcast secret message to the group. To overcome this limitation, this paper proposes an authenticated asymmetric GKA protocol. Instead of a common secret key, each group member negotiates a com- mon group public key and holds a different decryption key. The paper proposed protocol supports the dynamic nodes update of mobile networks, which has forward secrecy and backward secrecy of group key. This protocol is proven secure under the Bilinear Computational Diffie Hellman problem assumption and the performance analysis shows that the proposed scheme is highly efficient.     

A Modified Dynamic Differential Evolution Algorithm with Greedy Mutation Operation

Dynamic differential evolution （DDE） has been proposed to improve the performance of Differential evolution （DE） by dynamic updating of population. However, the convergence rate of DDE in optimizing a computationally expensive objective function still does not meet all our requirements. A new local search operation, greedy mutation operation, is proposed and embedded into DDE. The greedy mutation operation differs itself from the purely random mutation by considering fitness information. Modifications in mutation ensure that the fitness of base vector is superior to the average fitness of population, which leads to develop offspring that are more fit to survive than those generated from purely random operators. The modified DDE was tested against DE and DDE using five benchmark functions. The results show that the modified DDE converges faster without compromising solution quality.     

Layered Dynamic Schedulings for BP Decoding of LDPC Codes over GF（q）

A Layered dynamic scheduling （LDS） for Belief-propagation （BP） decoding of LDPC codes over GF（q） is presented, which is derived from the dynamic scheduling for the BP decoding of binary LDPC codes. In order to restrain the LDS from cycling in certain checknodes, a life-index for each check-node is adopted and the optimal value of the life-index is analyzed. Furthermore, in consideration of hardware implementation and decoding latency, a strategy, which allows many more checknodes to be updated in parallel, is introduced. Simulations show that the LDS with life-index speeds up the convergence rate and greatly improves the performance of the BP decoding at medium to high signal-to-noise ratio value, and the algorithm employing the LDS with life-index and the new strategy offers good trade-off between the performance and the decoding latency.     

Peak Power Minimizing Algorithm by Voltage Scheduling

For complicated electronic systems, to ensure high performance and reliability satisfaction, minimizing peak power consumption becomes one of the most important design goals. This paper addresses the problem of variable voltage scheduling on multlprocessor distributed systems, with the goal of shaping the power profile to minimizing peak power. A low peak power algorithm named LPPA is proposed to optimize power distribution via scaling voltage of the tasks on critical regions, based on the comprehensive analysis of how power consumption varies with latency. Compared with previous low peak power techniques, which simply scale voltage of tasks according to their timing critical degree, LPPA additionally take the power profile into count to further decrease the peak power. Experimental results show that the proposed voltage scheduling technique significantly improves the power characteristics over the existing power profile unaware scheduling technique. Meanwhile, energy consumption reduction is also obtained.     

Utilizing Sub-topic Units for Patent Prior-Art Search

One of the defining challenges in patent prior-art search is the problem of representing a long, technical document as a query. Previously work on this problem has concentrated on single query representations of the patent application. In the following paper, we describe an approach which uses multiple query representations generated from semantically coherent passages extracted from patent documents. We validate our technique in an experiment using the CLEF-IP 2011 patent search collection. Our system achieves statistically significant improvements over various state-of-art query generation techniques.     

Distributed In-Time Multiple Coverage Holes Healing Protocol

Coverage holes often appear in wireless sensor networks due to sensor failure or the inheritance of sensor＇s random distribution. In the hybrid model, mobile sensors in the network are acquired to heal coverage holes by their mobifity. When multiple coverage holes appear in the sensor network and each of them has a time requirement （in which the coverage hole has to be healed）, conflicts for the requests of the same mobile sensor may arise. A distributed multiple mobile sensor schedufing protocol （DMS） is proposed in this paper to solve this problem by finding mobile sensors in the time response zone defined by the time requirement of each coverage hole. Simulation results show that DMS can well schedule the mobile sensors to move to multiple coverage holes within the time requirement.     

Distributed Source Localization under Anchor Position Uncertainty

In this work, a distributed source position- ing approach is developed based on Alternating direction method of multipliers （ADMM）. First, a centralized posi- tioning method is developed under case of the anchor un- certainty. And then, the method is realized in a distributed way using ADMM. Simulation results show that the cen- tralized one is robust to the anchor errors and distributed one has similar performance as the centralized one.     

Performance Analysis Applications in Virtual of Interactive Desktop Machine Environment

To assess the feasibility of hosting interactive desktop applications in virtual machine environment, this paper measures and analyzes the interactive performance of the desktop operating system in Xen environment. The experimental results show that short-latency input events are not sensitive to the virtualization overhead but the average response latencies of long-latency events in Xen system are remarkably longer than the latencies in the unvirtualized system. In fact, the larger variance of response latency, instead of the longer latency itself, is the most important performance problem for the virtualized approach. Our analysis shows that the main reason of the poor performance is the insufficient CPU share of the driver-specific virtual machine, which causes the long freezing periods of other guest virtual machines. While the offline adjustment can alleviate the problem, an online adaptive adjustment of the CPU allocation is necessary for the practical use of Xen hosted desktop systems.     

A Key Pre-distribution Scheme for Wireless Sensor Networks Using Hexagonal Deployment knowledge

To insure security for the messages sent among sensor nodes, it is important to establish keys for encryption to nodes which have had no prior direct contact with each other in wireless sensor networks. To solve the problem above, several key pre-distribution schemes have been proposed. In this paper, we present a novel key pre-distribution scheme using hexagonal deployment knowledge, in which the q-composite keys scheme has been improved. By using hexagonal division over the deploy- ment region, the probability of sharing common keys be- tween each sub-region and its neighboring sub-regions are equal, so the keys for encryption can be generated efficiently. The analytic results show that the scheme we pre-sented can achieve a higher connection probability of any two neighboring nodes than previous schemes, and net- work resilience against node capture can be substantially improved.