- 高凤荣;邢春晓;杜小勇;王珊;
Personalized service systems are an effective way to help users obtain recommendations for unseen items, within the enormous volume of information available based on their preferences. The most commonly used personalized service system methods are collaborative filtering, content-based filtering, and hybrid filtering. Unfortunately, each method has its drawbacks. This paper proposes a new method which unified partition-based collaborative filtering and meta-information filtering. In partition-based collaborative filtering the user-item rating matrix can be partitioned into low-dimensional dense matrices using a matrix clustering algorithm. Recommendations are generated based on these low-dimensional matrices. Additionally, the very low ratings problem can be solved using meta-information filtering. The unified method is applied to a digital resource management system. The experimental results show the high efficiency and good performance of the new approach.
2007年01期 1-8页 [查看摘要][在线阅读][下载 232K] [下载次数：209 ] |[网刊下载次数：0 ] |[引用频次：26 ] |[阅读次数：0 ] - 刘丹;孙金玮;魏国;
This paper describes a data reconstruction technique for a multi-function sensor based on the M-estimator, which uses least squares and weighted least squares method. The algorithm has better robustness than conventional least squares which can amplify the errors of inaccurate data. The Mestimator places particular emphasis on reducing the effects of large data errors, which are further overcome by an iterative regression process which gives small weights to large off-group data errors and large weights to small data errors. Simulation results are consistent with the hypothesis with 81 groups of regression data having an average accuracy of 3.5%, which demonstrates that the M-estimator provides more accurate and reliable data reconstruction.
2007年01期 9-13页 [查看摘要][在线阅读][下载 314K] [下载次数：52 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ] - 胡亭亮;朱纪洪;孙增圻;
In this paper we present a robust adaptive control for a class of uncertain continuous time multiple input multiple output (MIMO) nonlinear systems. Multiple multi-layer neural networks are employed to approximate the uncertainty of the nonlinear functions, and robustifying control terms are used to compensate for approximation errors. All parameter adaptive laws and robustifying control terms are derived based on Lyapunov stability analysis so that, under appropriate assumptions, semi-global stability of the closed-loop system is guaranteed, and the tracking error asymptotically converges to zero. Simulations performed on a two-link robot manipulator illustrate the approach and its performance.
2007年01期 14-21页 [查看摘要][在线阅读][下载 411K] [下载次数：95 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ] - 谢翔;李国林;王志华;
This paper presents a novel method of lossy image compression for digital colorful image sensors with Bayer color filter arrays (CFAs) based on an analysis of existing compression-first methods. An improved compression and decompression structure is described. The captured CFA raw data are firstly lowpass filtered in RGB space by a smooth filter followed by a down-sampling operation. Next, the data are transformed from RGB space to YCbCr space. Lastly, the filtered data in YCbCr space are compressed directly before full color interpolation to avoid redundancy. The presented method can provide a lower compression ratio and has lower complexity than both conventional interpolation-first image compression methods and other similar existing compression-first methods.
2007年01期 22-29页 [查看摘要][在线阅读][下载 420K] [下载次数：106 ] |[网刊下载次数：0 ] |[引用频次：4 ] |[阅读次数：0 ] -
<正>Twenty-three university presidents from the U.S. and China discussed higher education issues such as the cultiva-tion of talent and international cooperation at the US-China University Presidents Roundtable on November 16,
2007年01期 29页 [查看摘要][在线阅读][下载 76K] [下载次数：22 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ] - 吕宁;王雄;
This paper presents some practical applications of signed directed graphs (SDGs) to computeraided hazard and operability study (HAZOP) and fault diagnosis, based on an analysis of the SDG theory. The SDG is modeled for the inversion of synthetic ammonia, which is highly dangerous in process industry, and HAZOP and fault diagnosis based on the SDG model are presented. A new reasoning method, whereby inverse inference is combined with forward inference, is presented to implement SDG fault diagnosis based on a breadth-first algorithm with consistency rules. Compared with conventional inference engines, this new method can better avoid qualitative spuriousness and combination explosion, and can deal with unobservable nodes in SDGs more effectively. Experimental results show the validity and advantages of the new SDG method.
2007年01期 30-37页 [查看摘要][在线阅读][下载 316K] [下载次数：540 ] |[网刊下载次数：0 ] |[引用频次：16 ] |[阅读次数：3 ] - 金亚;罗贵明;
The autoregressive moving average exogenous (ARMAX) model is commonly adopted for describing linear stochastic systems driven by colored noise. The model is a finite mixture with the ARMA component and external inputs. In this paper we focus on a parameter estimate of the ARMAX model. Classical modeling methods are usually based on the assumption that the driven noise in the moving average (MA) part has bounded variances, while in the model considered here the variances of noise may increase by a power of log n. The plant parameters are identified by the recursive stochastic gradient algorithm. The diminishing excitation technique and some results of martingale difference theory are adopted in order to prove the convergence of the identification. Finally, some simulations are given to show the reliability of the theoretical results.
2007年01期 38-43页 [查看摘要][在线阅读][下载 257K] [下载次数：71 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ] - 谢峰;陶霖密;徐光祐;
An algorithm is presented for estimating the direction and strength of point light with the strength of ambient illumination. Existing approaches evaluate these illumination parameters directly in the high dimensional image space, while we estimate the parameters in two steps: first by projecting the image to an orthogonal linear subspace based on spherical harmonic basis functions and then by calculating the parameters in the low dimensional subspace. The test results using the CMU PIE database and Yale Database B show the stability and effectiveness of the method. The resulting illumination information can be used to synthesize more realistic relighting images and to recognize objects under variable illumination.
2007年01期 44-50页 [查看摘要][在线阅读][下载 730K] [下载次数：33 ] |[网刊下载次数：0 ] |[引用频次：4 ] |[阅读次数：0 ] - 柳杨华;徐光祐;
Realistic personalized face animation mainly depends on a picture-perfect appearance and natural head rotation. This paper describes a face model for generation of novel view facial textures with various realistic expressions and poses. The model is achieved from corpora of a talking person using machine learning techniques. In face modeling, the facial texture variation is expressed by a multiview facial texture space model, with the facial shape variation represented by a compact 3-D point distribution model (PDM). The facial texture space and the shape space are connected by bridging 2-D mesh structures. Levenberg-Marquardt optimization is employed for fine model fitting. Animation trajectory is trained for smooth and continuous image sequences. The test results show that this approach can achieve a vivid talking face sequence in various views. Moreover, the animation complexity is significantly reduced by the vector representation.
2007年01期 51-57页 [查看摘要][在线阅读][下载 338K] [下载次数：69 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ] - 张超;周一鸣;张曾科;
The fractal image encoding method has received much attention for its many advantages over other methods, such as high decoding quality at high compression ratios. However, because every range block must be compared to all domain blocks in the codebook to find the best-matched one during the coding procedure, baseline fractal coding (BFC) is quite time consuming. To speed up fractal coding, a new fast fractal encoding algorithm is proposed. This algorithm aims at reducing the size of the search window during the domain-range matching process to minimize the computational cost. A new theorem presented in this paper shows that a special feature of the image can be used to do this work. Based on this theorem, the most inappropriate domain blocks, whose features are not similar to that of the given range block, are excluded before matching. Thus, the best-matched block can be captured much more quickly than in the BFC approach. The experimental results show that the runtime of the proposed method is reduced greatly compared to the BFC method. At the same time, the new algorithm also achieves high reconstructed image quality. In addition, the method can be incorporated with other fast algorithms to achieve better performance. Therefore, the proposed algorithm has a much better application potential than BFC.
2007年01期 58-62页 [查看摘要][在线阅读][下载 296K] [下载次数：63 ] |[网刊下载次数：0 ] |[引用频次：12 ] |[阅读次数：0 ] - 王云峰;边计年;洪先龙;周强;吴强;
As the feature size of integrated circuits is reduced to the deep sub-micron level or the nanometer level, the interconnect delay is becoming more and more important in determining the total delay of a circuit. Re-synthesis after floorplan is expected to be very helpful for reducing the interconnect delay of a circuit. In this paper, a force-balance-based re-synthesis algorithm for interconnect delay optimization after floorplan is proposed. The algorithm optimizes the interconnect delay by changing the operation scheduling and the functional unit allocation and binding. With this method the number and positions of all functional units are not changed, but some operations are allocated or bound to different units. Preliminary experimental results show that the interconnect wire delays are reduced efficiently without destroying the floorplan performance.
2007年01期 63-69页 [查看摘要][在线阅读][下载 267K] [下载次数：47 ] |[网刊下载次数：0 ] |[引用频次：1 ] |[阅读次数：0 ] - 陈文涛;金德鹏;曾烈光;
The Ethernet over E1 approach, which takes advantage of widely deployed telecom networks, is an efficient and economical way to interconnect two Ethernets in different regions. Two Ethernet over E1 schemes, namely a byte granularity scheme and a frame granularity scheme are discussed. The byte granularity scheme partitions Ethernet frames into several pieces for transmission and has a strict requirement on the maximum delay difference of multiple E1 links. To solve this problem, the newly proposed frame granularity scheme transmits separately each frame through E1 links without any partitioning. The architecture designs of both schemes are presented. This paper evaluates the throughput and delay performances of both schemes, both analytically from results calculated from delay models and using test results from field programmable gate array (FPGA) implementation. Although the frame granularity scheme has a slightly worse delay performance, it has a higher throughput, and is the only choice able to overcome large delay differences of the E1 links.
2007年01期 70-76页 [查看摘要][在线阅读][下载 205K] [下载次数：47 ] |[网刊下载次数：0 ] |[引用频次：1 ] |[阅读次数：0 ] - 马立伟;孙义和;
Network-on-chip (NoC) is a new design paradigm for system-on-chip intraconnections in the billion-transistor era. Application specific on-chip network design is essential for NoC success in this new era. This paper presents a class of source routing switches that can be used to efficiently form arbitrary network topologies and that can be optimized for various applications. Hardware description language versions of the networks can be generated automatically for simulations and for syntheses. A series of switches and networks has been configured with their performances including latency, delay, area, and power, and analyzed theoretically and experimentally. The results show that this NoC architecture provides a large design space for application specific on-chip network designs.
2007年01期 77-85页 [查看摘要][在线阅读][下载 476K] [下载次数：107 ] |[网刊下载次数：0 ] |[引用频次：1 ] |[阅读次数：0 ] - 赵安邦;王文生;崔慧娟;唐昆;
A simple, efficient multiple description coding (MDC) algorithm was developed based on weighted signal combinations. The scheme uses the standard video encoder with a pre-processing stage to generate multiple descriptions. The decoder then uses a post-processing algorithm to combine the descriptions to provide better image quality. A scalable codec in the MDC system allows the system to provide multiple descriptions and scalability at the same time. In addition, since the different scalable descriptions may have different qualities, a simple averaging process is not optimal. An optimal weighted combination of the two de- scriptions was developed based on the signal to noise ratios. Compared with the simple average combination, the algorithm significantly improved the video quality, especially with large quality differences between the two descriptions, with gains of up to 3.56 dB.
2007年01期 86-90页 [查看摘要][在线阅读][下载 158K] [下载次数：68 ] |[网刊下载次数：0 ] |[引用频次：1 ] |[阅读次数：0 ] -
<正>An inauguration ceremony was held to mark the opening of the Center for International Transboundary Water and Eco-Security, Tsinghua University (CITWES) on December 10, 2006. Zhai Haohui, Vice Minister of MWR,
2007年01期 90页 [查看摘要][在线阅读][下载 55K] [下载次数：20 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ] - 王小斌;唐万生;赵瑞清;
In economic order quantity models without backordering, both the stock cost of each unit quantity and the order cost of each cycle are characterized as independent fuzzy variables rather than fuzzy numbers as in previous studies. Based on an expected value criterion or a credibility criterion, a fuzzy expected value model and a fuzzy dependent chance programming (DCP) model are constructed. The purpose of the fuzzy expected value model is to find the optimal order quantity such that the fuzzy expected value of the total cost is minimal. The fuzzy DCP model is used to find the optimal order quantity for maximizing the credibility of an event such that the total cost in the planning periods does not exceed a certain budget level. Fuzzy simulations are designed to calculate the expected value of the fuzzy objective function and the credibility of each fuzzy event. A particle swarm optimization (PSO) algorithm based on a fuzzy simulation is designed, by integrating the fuzzy simulation and the PSO algorithm. Finally, a numerical example is given to illustrate the feasibility and validity of the proposed algorithm.
2007年01期 91-96页 [查看摘要][在线阅读][下载 226K] [下载次数：132 ] |[网刊下载次数：0 ] |[引用频次：9 ] |[阅读次数：0 ] - Kamran Ullah Khan;杨建;
Different methods proposed so far for accurate classification of land cover types in polarimetric synthetic aperture radar (SAR) image are data specific and no general method is available. A novel hybrid framework for this classification was developed in this work. A set of effective features derived from the coherence matrix of polarimetric SAR data was proposed. Constituents of the feature set are wavelet, texture, and nonlinear features. The proposed feature set has a strong discrimination power. A neural network was used as the classification engine in a unique way. By exploiting the speed of the conjugate gradient method and the convergence rate of the Levenberg-Marquardt method (near the optimal point), an overall speed up of the classification procedure was achieved. Principal component analysis (PCA) was used to shrink the dimension of the feature vector without sacrificing much of the classification accuracy. The proposed approach is compared with the maximum likelihood estimator (MLE) based on the complex Wishart distribution and the results show the superiority of the proposed method, with the average classification accuracy by the proposed method (95.4%) higher than that of the MLE (93.77%). Use of PCA to reduce the dimensionality of the feature vector helps reduce the memory requirements and computational cost, thereby enhancing the speed of the process.
2007年01期 97-104页 [查看摘要][在线阅读][下载 465K] [下载次数：80 ] |[网刊下载次数：0 ] |[引用频次：5 ] |[阅读次数：0 ] - 姜睿;罗贵明;
The least-squares (LS) algorithm has been used for system modeling for a long time. Without any excitation conditions, only the convergence rate of the common LS algorithm can be obtained. This paper analyzed the weighted least-squares (WLS) algorithm and described the good properties of the WLS algorithm. The WLS algorithm was then used for adaptive control of linear stochastic systems to show that the linear closed-loop system was globally stable and that the system identification was consistent. Compared to the past optimal adaptive controller, this controller does not impose restricted conditions on the coefficients of the system, such as knowing the first coefficient before the controller. Without any persistent excitation conditions, the analysis shows that, with the regulation of the adaptive control, the closed-loop system was globally stable and the adaptive controller converged to the one-step-ahead optimal controller in some sense.
2007年01期 105-110页 [查看摘要][在线阅读][下载 177K] [下载次数：65 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ] - 牛志升;李蕾;
Cochannel interference introduced by frequency reuse under space division multiple access can be reduced by controlling the frequency sharing according to the users' spatial separability. The system avoids placing highly correlated users in space into the same spatial multiplexing group. This paper describes a user spatial compatibility evaluation criterion based on channel matrix structure analysis. The interferences are evaluated using the distances between subspaces spanned by the users' channel matrixes. The analysis considers row space and column space. Simulation results show that the row space distance criterion significantly outperforms other criteria because it fully utilizes the information from the users' chan- nel matrixes. As the dimensions of the users' channel matrixes increase, the matrix structure analysis-based criteria become even more efficient.
2007年01期 111-114页 [查看摘要][在线阅读][下载 160K] [下载次数：70 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ] <正>Tsinghua Science and Technology (Tsinghua Sci Technol), a comprehensive academic journal sponsored by Tsinghua University, is published bimonthly. This journal aims at presenting the up-to-date scientific achievements
2007年01期 115页 [查看摘要][在线阅读][下载 534K] [下载次数：16 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：0 ]