ISSN 1000-1239 CN 11-1777/TP


    Default Latest Most Read
    Please wait a minute...
    For Selected: Toggle Thumbnails
    Recent Advances in Bayesian Machine Learning
    Zhu Jun,Hu Wenbo
    Journal of Computer Research and Development    2015, 52 (1): 16-26.   DOI: 10.7544/issn1000-1239.2015.20140107
    Abstract4486)   HTML128)    PDF (2137KB)(5021)       Save
    With the fast growth of big data, statistical machine learning has attracted tremendous attention from both industry and academia, with many successful applications in vision, speech, natural language, and biology. In particular, the last decades have seen the fast development of Bayesian machine learning, which is now representing a very important class of techniques. In this article, we provide an overview of the recent advances in Bayesian machine learning, including the basics of Bayesian machine learning theory and methods, nonparametric Bayesian methods and inference algorithms, and regularized Bayesian inference. Finally, we also highlight the challenges and recent progress on large-scale Bayesian learning for big data, and discuss on some future directions.
    Related Articles | Metrics
    Feature Selection for Multi-Label Classification Based on Neighborhood Rough Sets
    Duan Jie, Hu Qinghua,Zhang Lingjun,Qian Yuhua,Li Deyu
    Journal of Computer Research and Development    2015, 52 (1): 56-65.   DOI: 10.7544/issn1000-1239.2015.20140544
    Abstract2165)   HTML13)    PDF (3560KB)(1866)       Save
    Multi-label classification is a kind of complex decision making tasks, where one object may be assigned with more than one decision label. This kind of tasks widely exist in text categorization, image recognition, gene function analysis. Multi-label classification is usually described with high-dimensional vectors, and some of the features are superfluous and irrelevant. A great number of feature selection algorithms have been developed for single-label classification to conquer the curse of dimensionality. However, as to multi-label classification, fewer researches have been reported for designing feature selection algorithms. In this work, we introduce rough sets to multi-label classification for constructing a feature selection algorithm. We redefine the lower approximation and dependency, and discuss the properties of the model. After that, we design a neighborhood rough sets based feature selection algorithm for multi-label classification. Experimental results show the effectiveness of the proposed algorithm.
    Related Articles | Metrics
    History and Recent Developments of AVS Video Coding Standards
    Ma Siwei
    Journal of Computer Research and Development    2015, 52 (1): 27-37.   DOI: 10.7544/issn1000-1239.2015.20140106
    Abstract2029)   HTML12)    PDF (3233KB)(1012)       Save
    AVS(audio video coding standard) is the informal name of Work group for Digital Audiovideo Coding Standard of China, which was founded by the Science and Technology Department under former Ministry of Information Industry in June 2002, approved by Standardization Administration of China. The role of the group is to establish general technical standards for the compression, decoding, processing, and the representation of digital audio-video, thereby enabling digital audio-video equipment and systems with high-efficiency and economical coding/decoding technologies. After more than ten years, AVS has established a series of video coding standards, including AVS1, AVS+ and AVS2. AVS1 and AVS2 are named from the first and second stage work of AVS, or the first and second generation standard, and AVS+ was established specially for China high definition TV broadcasting specially. AVS1 and AVS+ have been finished and widely used in various applications so far, and AVS2 is still under developing and will be released soon. This paper provides an overview of the history and recent developments of AVS video coding standards, including the key tools used in AVS and the comparison with the state-of-the-art technology, e.g. HEVC/H265. Moreover, a brief discussion and conclusion on the future video coding are provided.
    Related Articles | Metrics
    Survey on Computer System Simulator
    Liu Yuchen,Wang Jia,Chen Yunji,Jiao Shuai
    Journal of Computer Research and Development    2015, 52 (1): 3-15.   DOI: 10.7544/issn1000-1239.2015.20140104
    Abstract1897)   HTML14)    PDF (1696KB)(1570)       Save
    Computer system simulator has long been a useful tool for researchers. It is applied in many different areas, from code design to software programming. In the development of simulators, performance has always been the main focus of researchers, and the improvement of performance will in return benefit the performance of real computers. A number of optimization work has been proposed in both serial and parallel simulations, such as threaded code, binary translation, FPGA accelerator, simulation separation techniques in serial simulation and the solution for the load balance, synchronization and communication in parallel simulation. In this paper, we provide several basic rules and structures that are used in common simulator design, and summarize recent studies of serial and parallel simulation and simulators. First, we introduce the current development of simulators, including current research results, technical problems and challenges. Then, we talk about the structure and the classification of current simulators. After that, the technique in serial simulators is introduced, and the optimization work in parallel simulation is also organized, according to the problems they tend to solve. Some mature simulators as well as simulation platforms are presented later in the paper. At last, potential issues and future work are also introduced.
    Reference | Related Articles | Metrics
    Xu Zhiwei
    Journal of Computer Research and Development    2015, 52 (1): 1-2.  
    Abstract1656)   HTML5)    PDF (307KB)(947)       Save
    Related Articles | Metrics
    Topology Optimization for Minimal Volume in 3D Printing
    Xu Wenpeng,Wang Weiming,Li Hang,Yang Zhouwang,Liu Xiuping,Liu Ligang
    Journal of Computer Research and Development    2015, 52 (1): 38-44.   DOI: 10.7544/issn1000-1239.2015.20140108
    Abstract1626)   HTML5)    PDF (2555KB)(937)       Save
    Compared with the traditional products manufacturing pattern, the cost of 3D printing products is still relatively high. It is important to optimize model to reduce print material consumption and printing costs without sacrificing print quality of the object surface. To solve this problem, we present a topology optimization algorithm for minimal volume in 3D printing with traditional evolutionary structural optimization methods combined with Von Mises stress. The algorithm calculates Von Mises stress of the model to guide the evolution of the volume reducing, until the maximum Von Mises stress reaches the allowable stress value of the material. Furthermore, we introduce multi-resolution technology to accelerate optimization computing from the coarse tetrahedral meshes to fine meshes, which effectively improves the computational efficiency. Compared with other existing methods, the optimization results of our method can be more flexible and better reflect the load transfer path of model under the given force.
    Related Articles | Metrics
    A Replay System for Performance Analysis of Multi-Threaded Programs
    Zheng Long, Liao Xiaofei, Wu Song, Jin Hai
    Journal of Computer Research and Development    2015, 52 (1): 45-55.   DOI: 10.7544/issn1000-1239.2015.20140105
    Abstract1431)   HTML1)    PDF (2461KB)(793)       Save
    In recent years, it is a hotspot for program analysis to detect performance bugs in multi-threaded applications. However, traditional record/replay systems focusing on concurrent anomalies have many limitations to tackle the issues of performance bugs, such as replay overhead and imprecision of replay-based execution time. To cope with the problems above, this paper proposes an improved replay system PerfPlay which can be used for the performance analysis of multi-threaded programs. To be specific, we first collect and analyze the requisite information for the program performance. Secondly, the different replay strategies are discussed and then we present a novel schedule-driven strategy to ensure the performance fidelity of replay system. Finally, we study the classical performance problem of “inter-thread unnecessary lock contention” under the framework of PerfPlay. Compared with the traditional replay strategies, our experimental results demonstrate the performance fidelity of PerfPlay. Through the case study, we find a few performance bugs in real-world and further verify the effectiveness of PerfPlay.
    Related Articles | Metrics