Spatial coherence of a high-flux, ultrashort-pulse, extreme-ultraviolet (EUV) source was discussed. A technique for measuring the spectrum of the EUV emission by deconvolution of a double-slit interference pattern was...
详细信息
ISBN:
(纸本)1557527067
Spatial coherence of a high-flux, ultrashort-pulse, extreme-ultraviolet (EUV) source was discussed. A technique for measuring the spectrum of the EUV emission by deconvolution of a double-slit interference pattern was presented. A large fraction of the 1 mm diameter EUV beam ∼95 cm from the output of the waveguide was sampled. Results showed that in each case, the fringe contrast in the region of spatial overlap between the two waves was 100%.
Now I/O is the bottleneck of performance in the comp.ter system. Although disk is the main second storage device, there is a performance gap between disk and RAM because of mechanism characters of the disk system. And...
详细信息
Now I/O is the bottleneck of performance in the comp.ter system. Although disk is the main second storage device, there is a performance gap between disk and RAM because of mechanism characters of the disk system. And it is very difficult to solve the reducing the latency of the synchronous operation. At the same time, synchronous operation of small data is general in I/O operation. Although most research focusing on how to improve the bandwidth of I/O, a method trading capacity for reducing latency of I/O is proposed. This method can insure data durability.
Text filtering involves feature extraction, text classification, and corpus selection. Both algorithms and architecture patterns affects the effectiveness of ultimate information filtering systems. To improve effectiv...
详细信息
Text filtering involves feature extraction, text classification, and corpus selection. Both algorithms and architecture patterns affects the effectiveness of ultimate information filtering systems. To improve effectiveness and reusability of information filtering system, a domain specific information filtering framework called IFDF (information filtering domain-specific framework) is constructed, which is represented in UML. Then IFDF is extended with event-based interface. Three architecture patterns, namely MAPAP (multi-algorithm parallel architecture pattern), MARAP (multi-algorithm refiner architecture pattern) and EDFAP (event-driven feedback architecture pattern) are proposed. They can also be used in other algorithm-intensive applications.
In order to specify real-time systems, many temporal logics such as Timed comp.tation Tree Logic, Metric Interval Temporal Logic and Real-Time Temporal Logic have been proposed. Although these logics are good at speci...
详细信息
In order to specify real-time systems, many temporal logics such as Timed comp.tation Tree Logic, Metric Interval Temporal Logic and Real-Time Temporal Logic have been proposed. Although these logics are good at specifying properties of real-time systems, they are not suitable for describing the implementations of such systems. Thus, the specifications and the implementations are usually described by different languages for real-time systems. A new linear temporal logic with clocks, called LTLC, is introduced. It is an extension of Manna and Pnueli's linear temporal logic. It can express both the properties and the implementations of real-time systems. With LTLC, systems can be described at many levels of abstraction, from high-level requirement specifications to low-level implementation models, and the conformance between different descriptions can be expressed by logical implication. This aspect of LTLC will be beneficial to the verification and the stepwise refinements of real-time systems.
An important step in text mining is to find a reasonable representation of the text. In the popular VSM (vector space module), where a text is represented as a vector, the coral problem is to term extraction, selectio...
详细信息
An important step in text mining is to find a reasonable representation of the text. In the popular VSM (vector space module), where a text is represented as a vector, the coral problem is to term extraction, selection and weighting. An iteration method is proposed to deal with the duplex phenomena found in term weighting and comp.te out the latent concept. Experimental results show that the latent concept could help to get better clustering results.
Based on BSP model and the concept of comp.tation-send segments, this paper proposes an asynchronous parallel comp.ting model, CSA-BSP, which can more accurately describe the performance parameters of parallel comp.te...
详细信息
Based on BSP model and the concept of comp.tation-send segments, this paper proposes an asynchronous parallel comp.ting model, CSA-BSP, which can more accurately describe the performance parameters of parallel comp.ters and guide programmers to write high efficient programs. This model utilizes the overlap of comp.tation and communication and makes communications spread around a super-step, which will reduce the congestion of communication in a traditional BSP super-step. Under CSA-BSP model, the execution time of a process can be estimated and its performance equation can be got. In this model, two processes can execute in different super steps, at most p-1 super steps away from each other. Using program is executing time as the parameter, authors analyze the efficiencies of parallel programs under BSP, A-BSP and CSA-BSP models. comp.red with the BSP and A-BSP programs, CSA-BSP programs are more efficient. The results are verified by the programs of the Red and Black method and the matrix multiplication. In examples, comp.red to BSP programs, the efficiencies of CSA-BSP programs increase by 20% and 37%. To analyze the throughput of CSA-BSP model, another parameter, the total time used by all the processes in one application (PTS) is proposed. The CSA-BSP program of Red and Black method can reduce the PTS time by 8% against.that in the BSP program. During this time all resources have been released and they can be used by other tasks. Theoretical analysis and experiment results show that CSA-BSP model can more accurately analyze the performance parameters of parallel comp.ters. Programming with CSA-BSP model can enhance the performance both from improving the program is efficiency and from increasing the throughput of comp.ter systems.
Firstly, this paper discusses the difference between Bayesian estimation and classical parameter estimation and denotes the fundamental principle for incorporating the prior knowledge in Bayesian learning. Then it pro...
详细信息
Firstly, this paper discusses the difference between Bayesian estimation and classical parameter estimation and denotes the fundamental principle for incorporating the prior knowledge in Bayesian learning. Then it provides the incremental Bayesian learning model. This model explains the Bayesian learning process that changes the belief with the prior knowledge and new examples information. By selecting the Dirichlet prior distribution, this process is shown in detail. For new examples for incremental learning, there exist two statuses: with labels and without labels. As for examples with labels, it is easy to update the classification parameter with the help of conjunct Dirichlet distribution. So it is the key point to learn from unlabeled examples. This paper gives a method measuring the classification loss with 0-1 loss. Meanwhile, to improve the algorithm performance, the pool-based technique is introduced. Because the basic operations in learning are updating the classification parameters and classifying test inst.nces incrementally, authors give their approximate expressions. For testing algorithm's efficiency, this paper makes an experiment on mushroom data set in UCI repository. The initial training set contains 6 labeled examples. Then several unlabeled examples are added. The final experimental results show that this algorithm is feasible and effective.
A theoretic model and method is proposed to decrease power consumption effectively by reducing execution frequency on multithreaded architecture. At first the comp.tation model is studied to recognize the thread which...
详细信息
A theoretic model and method is proposed to decrease power consumption effectively by reducing execution frequency on multithreaded architecture. At first the comp.tation model is studied to recognize the thread which can be executed at lower frequency, and the factor is comp.ted to slow down the frequency. Then, an algorithm and policy of comp.ler optimization combining with thread partition for low power is given based on the analysis of application program. This model and method can be used to exploit TLP (thread level parallelism) and decrease the power consumption effectively for the multithreaded multiprocessor architecture with scalable execution frequency.
A new E-Chunk based multi-engine machine translation model is proposed. The model is comp.sed of a head-driven lexicalized parser, a word-similarity based E-Chunk match engine and a bilingual E-Chunk based transfer en...
详细信息
A new E-Chunk based multi-engine machine translation model is proposed. The model is comp.sed of a head-driven lexicalized parser, a word-similarity based E-Chunk match engine and a bilingual E-Chunk based transfer engine. The optimal E-Chunk tiling is constructed in a bottom-up style efficiently. Preliminary experimental results show that it is effective in domain oriented machine translation.
Image segmentation is a well-known hard problem in image processing. In order to automatically extract discriminant regions from an image, this paper presents a novel method of region extraction, which is based on a S...
详细信息
Image segmentation is a well-known hard problem in image processing. In order to automatically extract discriminant regions from an image, this paper presents a novel method of region extraction, which is based on a SOM (self-organizing map) reduction algorithm presented in the paper. Firstly, according to a multi-feature extraction algorithm, the raw image is transformed into a feature map, in which each feature vector consists of three sub-features: 1) color feature-dominant color of a sub-region, 2) texture feature-MRSAR parameters of a sub-region, 3) and position feature-center coordinate of a sub-region. Secondly, SOM training algorithm is performed against.the feature map generated at the first step. A self-organizing map, in which the number of units is much smaller than that of feature vectors in the feature map, is created after SOM training. SOM training establishes a relationship between units in the SOM and feature vectors in the feature map. Those feature vectors, which are close with each other at the feature space, may map to the same unit of the SOM. Then, a family of reduced self-organizing maps is produced using a two-phase reduction algorithm of SOM. At the first phase, the unit, which has the least number of feature vectors at the map, will be reduced. At the second phase, two units, which are nearest at the feature space, will be merged. Those feature vectors mapping to the reduced unit will re-map to other neighbouring units according to a BMU match rule. Finally, in order to select an optimum one from a series of reduced self-organizing maps, an unsupervised cluster-validity analysis is performed. Pixels in the raw image can be grouped into different discriminant regions according to the relationship between the relevant feature map and the optimum reduced self-organizing map. At last, two evaluation experiments are given to verify the effectiveness of the new method.
暂无评论