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....

2017 Vol.30 Issue.12, Published 2017-12-30 Orignal Article        Orignal Article 1057 Parallel Gout: An ACP-Based System Framework for Gout Diagnosis and Treatment WANG Fei-Yue, LI Changgui, GUO Yuanyuan, WANG Jing, WANG Xiao, QIU Tianyu, Meng Xiangbing, SHI Xiaobo In order to improve the accuracy and the efficiency of diagnosis and treatment of Gout in complicated situation and break through the gap between doctor′s profession, an artificial societies, computational experiments, and parallel execution(ACP) based parallel gout diagnosis and treatment framework is presented, named Parallel Gout. Parallel Gout could construct an artificial gout diagnosis and treatment system to represent and simulate the real procedure of medical diagnosis and treatment system. It can train and evaluate the various models of diagnosis and treatment of gout by Computational Experiments, and to manage and control the real medical system by Parallel Execution, to achieve the automatic and intelligent diagnosis and treatment procedure for gout. Parallel Gout can help doctors to reduce the likelihood of misdiagnosis and therapeutic mistakes, and increase the efficiency of medical procedure as well improve the professionality of doctors. Parallel Gout can also help the patients to manage the chronic diseases and keep healthy. The application of Parallel Gout in the diagnosis and treatment of gout is of great practical significance. This is an effective approach to motivate the traditional medical model to become intelligent and parallel, and contribute to a higher level of national health. 2017 Vol. 30 (12): 1057-1068 [Abstract] ( 580 ) [HTML 1KB] [ PDF 2584KB] ( 693 ) 1069 Decomposition Multi-objective Evolutionary Algorithm Based on Differentiated Neighborhood Strategy WANG Liping, WU Feng, ZHANG Mengzi, QIU Feiyue The performance of the decomposition-based multi-objective evolutionary algorithm is easily affected by the neighborhood of a subproblem. When the neighborhood is too large, the new solutions generated by the population propagation deviate from the Pareto set and the frequency of comparison between new solutions and old solutions in the neighborhood is increased for the updating subproblems. Consequently, the computational complexity of the algorithm is increased. If the neighborhood is too small, the algorithm easily falls into the local optimum. To solve this problem, a decomposed multi-objective evolutionary algorithm based on differentiated neighborhood strategy(MOEA/D-DN) is proposed. The suitable parameters are selected by analyzing the influence of different neighborhood sizes on the algorithm performance and different sizes of neighborhood for each subproblem are set according to the angle between their weight vectors and the central vector. Thus, the resource of algorithm is allocated more rationally and the velocity of searching for the optimal solution is also improved. Finally, the experimental results on the test functions of 2 dimensional ZDT and 3-5 dimensional DTLZ show that the convergence rate and the performance of MOEA/D-DN algorithm are improved obviously and the computational resource allocation of the algorithm is more reasonable. The overall solution quality is better. 2017 Vol. 30 (12): 1069-1082 [Abstract] ( 625 ) [HTML 1KB] [ PDF 1486KB] ( 396 ) 1083 Fast Learn++.NSE Algorithm Based on Sliding Window SHEN Yan, ZHU Yuquan, SONG Xinping The vote weight of each base-classifier in Learn++.NSE depends on all the error rates in the environments experienced, and the classification learning efficiency of the Learn++.NSE needs to be improved. Therefore, a fast Learn++.NSE algorithm based on sliding window(SW-Learn++.NSE) is presented in this paper. The sliding window is utilized to optimize the calculation of the weight. By only using the recent classification error rates of each base-classifier inside the sliding window to compute the vote weight, the SW-Learn++.NSE improves the efficiency of ensemble classification learning greatly. The experiment shows that the SW-Learn++.NSE achieves a higher execution efficiency with an equivalent classification accuracy compared to the Learn++.NSE. 2017 Vol. 30 (12): 1083-1090 [Abstract] ( 653 ) [HTML 1KB] [ PDF 955KB] ( 353 ) 1091 3D Object Recognition via Convolutional-Recursive Neural Network and Kernel Extreme Learning Machine LIU Yangyang, ZHANG Jun, GAO Xinjian, ZHANG Xudong, GAO Jun To tackle the issues of depth quality and non-linear classification in the large-scale RGB-D dataset, a 3D object recognition method is designed on the basis of convolutional-recursive neural network(CNN-RNN) and kernel extreme learning machine(KELM). Firstly, a depth coding algorithm is introduced to correct the numerical losses and noises in the original depth cue and unify the point cloud into the standard angle. And the original depth and the encoded depth are fused as the new depth cue. Secondly,multi-cue hierarchical features are learned using CNN-RNN. Meanwhile, the two-way spatial pyramid pooling method is exploited for each cue. Finally, KELM is constructed as the classifier to recognize 3D objects. The experimental results demonstrate the proposed method effectively improves the 3D object recognition accuracy and the classification efficiency. 2017 Vol. 30 (12): 1091-1099 [Abstract] ( 500 ) [HTML 1KB] [ PDF 1607KB] ( 470 ) 1100 Multi-semantic Metapath Based Classification Method in Heterogeneous Information Network DU Yongping, LIU Jingxuan, ZHANG Jinli Heterogeneous information network(HIN) is a kind of large-scale network containing many types of objects and complex links. The metapath based object classification method in HIN is proposed in this paper. The correlation feature matrix between nodes is built by the use of the metapath with different semantic information. In addition, the jumping path is extended to solve the problem of information sparseness. The experiments are conducted on DBLP dataset and the results show high performance of the proposed method in the complex network by using fewer labeled data. Furthermore, t-test result denotes that the performance is improved significantly by jumping path with small labeled data. 2017 Vol. 30 (12): 1100-1107 [Abstract] ( 629 ) [HTML 1KB] [ PDF 871KB] ( 436 ) 1108 Multi-label Contractive Hashing Method for Face Attributes Retrieval ZHAO Xuan, TAN Xiaoyang, SONG Ge Hashing methods possess advantages of low storage cost and fast query speed. However, most of the current hashing methods are designed to handle simple binary similarity rather than the complex multilevel semantic structure of the images associated with multiple labels. In this paper, a multi-label contractive hashing method(MLCH) is proposed to preserve the multilevel semantic similarity of multi-label images. In particular, the supervising information of attributes is proposed to help the training of the model and adopt an optimized selection algorithm to select training samples. Meanwhile, a contractive constrain term is added to the loss function to improve the quality of the generated binary codes. The proposed approach is evaluated on CelebA and PubFig databases, and the experimental results demonstrate its superiority over several state-of-the-art hashing methods on the task of large-scale image retrieval. 2017 Vol. 30 (12): 1108-1113 [Abstract] ( 321 ) [HTML 1KB] [ PDF 659KB] ( 299 ) 1114 Ensemble Face Pairs Distance Metric Learning for Cross-Age Face Verification WU Jiaqi, JING Liping Aiming at the variations of face pairs caused by different age gaps, an ensemble face pairs distance metric learning method(EFPML) is proposed for cross-age face verification. Firstly, the whole dataset is divided into several subsets with different age gaps. Then, a distance metric is learned for each subset. Finally, all face pairs are re-represented for many times via learnt distance metrics, the new representations are more distinguishable and the limited cross-age face data are expanded. To evaluate the proposed method, a series of experiments are conducted on two real-world cross age datasets, FG-NET and CACD. The results show that EFPML consistently outperforms the state-of-the-art methods and it has ability to reduce the effect of aging and improve verification performance. 2017 Vol. 30 (12): 1114-1120 [Abstract] ( 469 ) [HTML 1KB] [ PDF 916KB] ( 337 ) 1121 Feature Selection Algorithm Based on Joint Spectral Clustering and Neighborhood Mutual Information HU Minjie, ZHENG Liping, TANG Li, LIN Yaojin Aiming at some potential correlation between features in feature space, spectral clustering and neighborhood mutual information are exploited to explore the correlation features and obtain maximal relevant feature subset, respectively. And a feature selection algorithm combining spectral clustering and neighborhood mutual information is proposed. In this paper, the neighborhood mutual information is firstly applied to remove uncorrelated features, and then the spectral clustering is utilized to group features. The features of the same group are strongly correlated and the features of different groups are strongly different. Then, the feature subset strongly associated with class label is selected from each feature group. Finally, all selected feature subsets are collected together to form the final selected features. Extensive experiment is conducted with two different classifiers. Experimental results show that the proposed model effectively improves the classification performance with less features. 2017 Vol. 30 (12): 1121-1129 [Abstract] ( 578 ) [HTML 1KB] [ PDF 815KB] ( 349 ) 1130 Entity Disambiguation in Specific Domains Combining Word Vector and Topic Models MA Xiaojun, GUO Jianyi, WANG Hongbin, ZHANG Zhikun, XIAN Yantuan, YU Zhengtao When the Skip-gram word vector model deals with the polysemous words, only one word vector with mixed multiple semantics can be computed and different meanings of polysemous words can not be distinguished. In this paper, an entity disambiguation method combining the word vector and the topic model in specific domains is proposed. The word vector method is used to obtain the vector form of the reference term and the candidate entity from the background text and the knowledge base, respectively. The similarities of the context and the category reference are calculated, and the LDA topic model and the Skip-gram word vector models are used to obtain the word vector representation of different meanings of the polysemous words. Meanwhile, the domain keywords are extracted and then the domain topic keyword similarity are calculated. Finally, three types of features are combined, and the candidate entity with the highest similarity is selected as the final target entity. Experiments show that the proposed method has better disambiguation results than the existing disambiguation methods. 2017 Vol. 30 (12): 1130-1137 [Abstract] ( 661 ) [HTML 1KB] [ PDF 920KB] ( 503 ) 1138 Semi-supervised Labeled Hierarchical Dirichlet Process Topic Model for Document Categorization LI Yongzhong, ZHENG Tao The optimal structure of theme set can be automatically learned from the data with Hierarchical Dirichlet Process(HDP) topic model. However, the set of topics can not meet the semantic requirement. And in some theme models with labels it is difficult to set the parameters. Therefore, based on the known semantic labels and the certitude degree of labels, a semi-supervised labeled HDP topic model(SLHDP) and the accuracy evaluation index of random cluster are proposed in this paper. Higher weight is given by the known semantic labels. Combined with the property of the finite space being divided infinitely in Dirichlet process, the model is built via Chinese restaurant process. The experimental results on several Chinese and English datasets show that SLHDP model makes the topic set more reasonable in the text classification of large scale datasets. 2017 Vol. 30 (12): 1138-1148 [Abstract] ( 766 ) [HTML 1KB] [ PDF 579KB] ( 459 )

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