Home      About Journal      Editorial Board      Instructions      Ethics Statement      Contact Us                   中文

Judgement and Disposal of Academic Misconduct Article Copyright Transfer Agreement Proof of Confidentiality Requirements for Electronic Version More.... Chinese Association of Automation National ResearchCenter for Intelligent Computing System Institute of Intelligent Machines,Chinese Academy of Sciences More....

2013 Vol.26 Issue.7, Published 2013-07-30 Orignal Article        Orignal Article 609 The Concept of Max-Flow and Its Properties in Dynamic Networks ZHANG Ling Based on the dynamic quotient space model,the definitions of the max-flow and min-cut,and the condition that min-cut theorem holds in dynamic networks,are investigated. The analysis of the characteristics of max-flow in dynamic networks shows if the max-flow concept in the static environment is simply transferred to the dynamic one,the transformed max-flow does not have the two main properties,i.e. the additivity and total flow maximization. By introducing the concept of t-cut networks,a dynamic network is transformed into the combination of a set of static networks. Therefore,a new definition of the max-flow (steepest flow) is proposed and the defined concept is proved to have the above two properties. Then,a corresponding min-cut concept is presented and it is proved that the min-cut theorem holds under the new concepts of max-flow and min-cut. Finally,the algorithm for computing dynamic max-flow (steepest flow) is given. 2013 Vol. 26 (7): 609-614 [Abstract] ( 704 ) [HTML 0KB] [ PDF 379KB] ( 1034 ) 615 Research on the Model of Emotional Graded BDI Agents ZHANG Xiao-Jun,ZHOU Chang-Le In recent years,the Belief-Desire-Intention model is one of the most influential theories with respect to agent technology. Propositional dynamic logic and the infinite-valued ukasiewicz logic are blended to formulize the emotional graded BDI model. In order to represent the uncertainty behaviors by belief,desire,intention,fear,anxiety and self-confidence degree,the corresponding axioms are added to the ukasiewicz logic. The proposed emotional graded BDI agent model explicitly represents the uncertainty of mental attitudes and emotional states. The behavior of emotional graded BDI agent is determined by the different measure of each context which is added by concrete conditions. The model of beliefs,desires,intentions,fear,anxiety and self-confidence are axiomatized,and how they influence the behavior of agent is shown. This model is general enough to specify different types of agents. On the basis of presenting the language and semantics for this model,the axioms and rules are proposed for the emotional graded BDI logic and the dependability and completeness of the logic are proved. After illustrating the interrelations between different contexts for the model,an application of an emotional graded BDI agent for purchasing houses is given. Aiming at the representation and reasoning for the uncertainty,a formal support for distributed artificial intelligence is provided. 2013 Vol. 26 (7): 615-622 [Abstract] ( 568 ) [HTML 0KB] [ PDF 414KB] ( 550 ) 623 Adaptive Classification Algorithm for Gradual Concept-Drifting Data Zhang Jing-Xiang,Wang Shi-Tong,Deng Zhao-Hong At present,the concept-drifting phenomena in various datasets receives considerable attention. Aiming at the classification of concept drift,an adaptive neighbor projection mean discrepancy support vector machine (NMD-SVM) is proposed. The concept of the neighbor projection mean discrepancy in the reproducing kernel Hilbert space is defined to measure the discrepancy between adjacent sub-classifiers,and the distribution characteristics of data are integrated into the process of global optimization. Thus,the adaptability of classification algorithm is enhanced. The experimental results on the simulation and real datasets validate the effectiveness of the proposed algorithm. 2013 Vol. 26 (7): 623-633 [Abstract] ( 535 ) [HTML 0KB] [ PDF 985KB] ( 553 ) 634 Backward Cloud Transformation Algorithm for Realizing Stability Bidirectional Cognitive Mapping XU Chang-Lin,WANG Guo-Yin The uncertainty of human cognition for the objective world is mainly reflected by the basic unit of cognition,that is,concept. In this paper,based on the cloud model theory,a new stable backward cloud transformation algorithm is proposed firstly,and the stability of the method is illustrated by experimental comparison. Secondly,according to the characteristics of the transformation between the intension and extension of a concept,which is implemented by the forward cloud transformation and backward cloud transformation respectively,the stable bidirectional cognitive processes between intension and extension of concept are simulated by the forward cloud transformation algorithm and the backward cloud transformation algorithm. 2013 Vol. 26 (7): 634-642 [Abstract] ( 367 ) [HTML 0KB] [ PDF 633KB] ( 748 ) 643 Liquid State Machine Based Music Chord Sequence Recognition Algorithm ZHANG Guan-Yuan,WANG Bin A chord sequence recognition algorithm based on Liquid State Machine (LSM) is presented. Firstly,the music signal is segmented and Pitch Class Profile feature is extracted for every frame. Then,a LSM model is achieved after training. Two kinds of Bizarre Chord,chord appears probability vector and chord transformation matrix,are presented to post-process the chord sequence outputted by LSM. 8 sets of experimental data from neural network model,hidden Markov mode,echo state network model and LSM model show that the LSM gets a good performance,and the post-processing method also effectively improves the recognition accuracy. 2013 Vol. 26 (7): 643-647 [Abstract] ( 717 ) [HTML 0KB] [ PDF 507KB] ( 1193 ) 648 Overlapping Community Detection in Complex Networks Based on Cluster Prototypes JIANG Ya-Wen,JIA Cai-Yan,YU Jian Community structure is one of the important topological characteristics in complex networks. In real world,community structures in networks are often overlapped. And it is difficult to efficiently detect overlapping communities in a network. Optimizing Qov function directly is a solution for overlapping community detection,however,it is easy to generate a local optimal solution. To solve this problem,the concept of vertex central membership measure is introduced,and based on cluster prototypes of nodes in a network,an efficient framework is proposed to identify overlapping communities. Then the framework is applied to some classic clustering algorithms. The experimental results show that the proposed method avoids generating local optimal solution,and it is more efficient than the other algorithms on synthetic and real-world networks. 2013 Vol. 26 (7): 648-659 [Abstract] ( 533 ) [HTML 0KB] [ PDF 1582KB] ( 1338 ) 660 Kernel Canonical Correlation Analysis with Sparse Representation for Facial Expression Recognition ZHOU Xiao-Yan,ZHENG Wen-Ming,XIN Ming-Hai In facial expression recognition,the existences of image noises and the irrelevant image information to the expression changes usually influence the recognition accuracy. The traditional facial expression recognition method using kernel canonical correlation analysis (KCCA) is difficulty to solve this problem. To overcome this drawback,a kernel canonical correlation analysis with sparse representation (SKCCA) is proposed and applied to the facial expression recognition. The basic idea of the SKCCA method is to utilize the sparse representation approach to choose the spectral components of the facial feature matrix before modeling the correlation between facial feature matrix and the expression semantic feature matrix. Then,the expression recognition is carried out based on the correlation model. To demonstrate the superiority of the proposed method over the traditional KCCA method,extensive experiments are conducted on the JAFFE database and the experimental results confirm the effectiveness of the proposed method. 2013 Vol. 26 (7): 660-666 [Abstract] ( 601 ) [HTML 0KB] [ PDF 501KB] ( 741 ) 667 Web Content Extraction Based on Text Density Model ZHU Ze-De,LI Miao,ZHANG Jian,CHEN Lei,ZENG Xin-Hua In order to obtain useful content encompassed by a large number of irrelevant information,the content extraction becomes indispensable for web data application. An approach of web content extraction based on the text density model is proposed,which integrates page structure features with language features to convert text lines of page document into a positive or negative density sequence. Additionally,the Gaussian smoothing technique is used to revise the density sequence,which takes the content continuity of adjacent lines into consideration. Finally,the improved maximum sequence segmentation is adopted to split the sequence and extract web content. Without any human intervention or repeated trainings,this approach maintains the integrity of content and eliminates noise disturbance. The experimental results indicate that the web content extraction based on the text density model is widely adapted to different data sources,and both accuracy and recall rate of the proposed approach are better than those existing statistical models. 2013 Vol. 26 (7): 667-672 [Abstract] ( 598 ) [HTML 0KB] [ PDF 417KB] ( 1208 ) 673 Kernel Orthogonal Discriminant Local Tangent Space Alignment Algorithm ZHENG Gang-Min,XIA Su-Na,MA Yuan-Yuan,MA Xiao-Hu To address the drawbacks of the local tangent space alignment algorithm,a feature extraction method based on kernel transformation,kernel orthogonal discriminant local tangent space alignment algorithm (KOTSDA),is proposed. Firstly,the kernel mapping is performed to map the face data into a high dimensional nonlinear space and extract the nonlinear information.Then,tangent space discriminant analysis algorithm is used to preserve the intra-class local geometric structures and simultaneously maximize the inter-class difference in target function. Finally,KOTSDA is obtained with orthogonal constraints. It effectively avoids losing discriminant information which does not need to preprocess by PCA dimensional reduction. The experiments on ORL and Yale face databases demonstrate the effectiveness of the proposed algorithm. 2013 Vol. 26 (7): 673-679 [Abstract] ( 562 ) [HTML 0KB] [ PDF 485KB] ( 621 ) 680 Object Tracking Method Based on Particle Filter and Sparse Representation YANG Da-Wei,CONG Yang,TANG Yan-Dong Aiming at the problem of illumination variation in the object tracking of video image sequence,an object tracking method which uses sparse representation in particle filter frame is proposed based on LBP textual feature of object. The tracking particles of the current frame are generated by the last tracking result according to Gaussian distribution,the sparse representation of each particle to the template subspace is obtained by solving the l1-regularized least squares problem,and the tracking object is determined. Then,particle filter is used to propagate sample distribution in next tracking frame. In the procedure,the template is updated using a new template updating strategy. The experimental results validate the performance and advancement of the proposed method. 2013 Vol. 26 (7): 680-687 [Abstract] ( 662 ) [HTML 0KB] [ PDF 1843KB] ( 989 ) 688 Artificial Bee Colony Algorithm with Tracking Search and Immune Selection FU Li,LUO Jun To overcome the disadvantages of food source updating and the mechanism for onlooker bees to select food source in the Artificial Bee Colony (ABC) algorithm,an ABC algorithm based on tracking search and immune selection is proposed. The search methods for tracking the global optimal solution and randomly selecting solution are introduced on the basis of the original solution searching method. The searched optimal solution is selected as the candidate in order to accelerate the convergence of the population and improve the convergence of the algorithm. For the procedure of the onlooker bees selecting the food source,the regulation mechanism of antibody density in the immune system is introduced to keep the diversity of the population and enhance the global search ability of the traditional algorithm. The simulation results for 6 classical benchmark functions show that the improved algorithm has obvious advantages in the optimization accuracy and convergence rate compared with the original ABC,GABC,RABC and TABC. 2013 Vol. 26 (7): 688-694 [Abstract] ( 491 ) [HTML 0KB] [ PDF 636KB] ( 1086 ) 695 Top-k Query Calculations on Uncertain Dataset under MapReduce Framework LU Xin,CHEN Hua-Hui,DONG Yi-Hong,QIAN Jiang-Bo Top-k query is commonly used in the management and application on uncertain data. And the Top-k query semantics base on parameterized ranking functions (PRF) is the unified approach of various query semantics proposed in recent years. Aiming at the massive uncertain dataset,an effective method for the Top-k query based on MapReduce is proposed. Through the analysis on the Top-k query semantics of parameterized ranking functions,an algorithm is presented to get the upper bound of an un-retrieved tuple. In this way,the pruning strategy is used to get the Top-k tuples without retrieving every tuple in the dataset. Furthermore,two different strategies are presented to implement the proposed algorithm under the MapReduce computing model in Hadoop. Finally,two groups of experiments are performed aiming at a single-machine environment and the Hadoop distributed computing platform. The experimental results show that the proposed algorithm is more effective to deal with the Top-k queries for the massive uncertain data on running time. 2013 Vol. 26 (7): 695-700 [Abstract] ( 339 ) [HTML 0KB] [ PDF 698KB] ( 580 )

Supervised by China Association for Science and Technology Sponsored by Chinese Association of Automation NationalResearchCenter for Intelligent Computing System Institute of Intelligent Machines, Chinese Academy of Sciences Published by Science Press

Copyright © 2010 Editorial Office of Pattern Recognition and Artificial Intelligence Address: No.350 Shushanhu Road, Hefei, Anhui Province, P.R. China Tel: 0551-65591176 Fax:0551-65591176 Email: bjb@iim.ac.cn Supported by Beijing Magtech 　Email:support@magtech.com.cn