Syllable language models for Mandarin speech recognition: exploiting character language models.

Mandarin Chinese is based on characters which are syllabic in nature and morphological in meaning. All spoken languages have syllabiotactic rules which govern the construction of syllables and their allowed sequences. These constraints are not as restrictive as those learned from word sequences, but they can provide additional useful linguistic information. Hence, it is possible to improve speech recognition performance by appropriately combining these two types of constraints. For the Chinese language considered in this paper, character level language models (LMs) can be used as a first level approximation to allowed syllable sequences. To test this idea, word and character level n-gram LMs were trained on 2.8 billion words (equivalent to 4.3 billion characters) of texts from a wide collection of text sources. Both hypothesis and model based combination techniques were investigated to combine word and character level LMs. Significant character error rate reductions up to 7.3% relative were obtained on a state-of-the-art Mandarin Chinese broadcast audio recognition task using an adapted history dependent multi-level LM that performs a log-linearly combination of character and word level LMs. This supports the hypothesis that character or syllable sequence models are useful for improving Mandarin speech recognition performance.

Effects of cooling time on a closed LWR fuel cycle

In this study, the effects of cooling time prior to reprocessing spent LWR fuel has on the reactor physics characteristics of a PWR fully loaded with homogeneously mixed U-Pu or U-TRU oxide (MOX) fuel is examined. A reactor physics analysis was completed using the CASM04e code. A void reactivity feedback coefficient analysis was also completed for an infinite lattice of fresh fuel assemblies. Some useful conclusions can be made regarding the effect that cooling time prior to reprocessing spent LWR fuel has on a closed homogeneous MOX fuel cycle. The computational analysis shows that it is more neutronically efficient to reprocess cooled spent fuel into homogeneous MOX fuel rods earlier rather than later as the fissile fuel content decreases with time. Also, the number of spent fuel rods needed to fabricate one MOX fuel rod increases as cooling time increases. In the case of TRU MOX fuel, with time, there is an economic tradeoff between fuel handling difficulty and higher throughput of fuel to be reprocessed. The void coefficient analysis shows that the void coefficient becomes progressively more restrictive on fuel Pu content with increasing spent fuel cooling time before reprocessing.

On the representation of an inclined Gaussian beam

In most contemporary optics courses, Gaussian beams are demonstrated in the form of propagation along one coordinate axis. This is referred to as the conventional representation and is in fact a special form. In this paper, we derive the general representation of a Gaussian beam propagating obliquely to the coordinate axis, by performing a coordinate rotation transformation on the conventional representation. When doing so on the beam parameters, a restrictive condition has to be taken into account. Without this condition, the expressions for the beam parameters after the rotation are not consistent with the conventional ones.

秦巴山区中山地小流域土地持续利用模式探讨——以柞水县薛家沟流域为例

在分析薛家沟流域土地利用现状、持续利用条件和限制因素的基础上 ,依据流域立体分异特性 ,提出川平地以粮食种植、坡地以经济果木和防护林为主导的山地林果药菌立体开发的土地利用模式

沙产业发展现状及前景展望

沙产业是开发利用沙区地下、地面和地上的多种资源 ,创建沙地农业型和非农业型产业 ,形成产品系列生产 ,以获取持续规模的效益。制约沙产业发展的瓶颈因素是水。我国对沙区资源的开发利用已扩大到多个领域 ,沙区用材林基地建设、泥炭资源开发利用、热管技术应用、石油天然气开采等都初具规模。 2 1世纪将是沙产业崛起的世纪 ,发展沙产业前景广阔。

基于势场法的移动机器人路径规划仿真研究

针对势场法的障碍物附近目标不可达(GNRON)问题,采用改进斥力势场函数,把机器人和目标的相对距离考虑进去,从而确保目标点为整个势场的全局最小点,使得机器人能够顺利到达目标。针对局部极小引起的陷阱区域问题,提出了增加引导点的方法,使得机器人能够快速走出陷阱区域,向目标点移动。通过仿真实验,还实现了机器人在限定区域内漫游。改进后的势场法适用于复杂环境下的移动机器人路径规划。仿真结果证明了此方法的有效性。

可重构星球探测机器人的运动学建模及轨迹规划

提出了可重构星球探测机器人的概念,对系统中子机器人的研究进行了重点论述.通过设计恰当的子机器人连杆坐标系,利用Denavit Hartenberg方法完成了子机器人的运动学建模,并直接给出了子机器人的运动学正解模型.由于使用单一的求解算法不能求出工作空间的封闭解,因此综合利用代数法、几何法原理及空间投影关系,结合子机器人的结构特殊性推导出了运动学逆解,从而得到了工作空间内的所有解.在此基础上,考虑结构间的约束关系,给出了子机器人的工作空间及轨迹规划方法.最后,使用OpenGL对设计的子机器人系统进行了运动学仿真实验,实验以末端操作器的直线运动为例,充分考虑空间几何的关系,其结果有效地证明了建模及轨迹规划的正确性.

目标位姿测量中的三维视觉方法

要测量出一组特征点分别在两个空间坐标系下的坐标 ,就可以求解两个空间目标间的位姿关系 .实现上述目标位姿测量方法的前提条件是要保证该组特征点在不同坐标系下 ,其位置关系相同 ,但计算误差的存在却破坏了这种固定的位置关系 .为此 ,提出了两种基于模型的三维视觉方法——基于模型的单目视觉和基于模型的双目视觉 ,前者从视觉计算的物理意义入手 ,通过简单的约束迭代求解实现模型约束 ;后者则将简单的约束最小二乘法和基于模型的单目视觉方法融合在一起来实现模型约束 .引入模型约束后 ,单目视觉方法可以达到很高的测量精度 .而基于模型的双目视觉较传统的无模型立体视觉方法位移精度提高有限 ,但姿态精度提高很多

三峡库区巴东组泥灰岩溶蚀作用与岩体力学性质弱化研究

Abstract The karsrt erosion engineering geology became a highlight problem in recent years, in particularly, the karst erosion of marlite of Badong formation made the rock mechanics weaken in Three Gorges Reservoir area, which reduces the safety of slope. During the immigrant construction, many high slopes have been formed, whose instabilities problems pose serious threats to the safety of the people and properties. The accidents of the slope failure take place now and then. By testing, it has been found that the karst erosion pattern and dissolution rate of marlite are not weaker than that of the pure limestone. Furthermore, owning to the weathering and unloading, the karst erosion of the marlite will reach certain depth of the slope, which is named infiltrated karst erosion. The karst erosion made the rock mass quality of slope or foundation worse in a large scale. The karst erosion geological disasters, taken place or not, has become the main restrictive factors to the social stability and economic development. Thus the karst erosion process and mechanism of marlite of Badong formation are studied as the main content of this dissertation. The weakening characteristic of rock mass mechanics parameters are studied along with the rock mass structure deformation and failure processes in the course of the karst erosion. At first, the conditions and influencing factors of the karst erosion are analyzed in the investigative region, on the basis of different karst erosion phenomenon of the marlite and different failure modes of slope. Then via indoor the karst erosion tests, it is analyzed that the karst erosion will change the rock mass composition and its structure. Through test, the different karst erosion phenomena between micro and macro have been observed, and the karst erosion mechanism of the marlite has been summarized. Damage theory is introduced to explain the feature of dissolution pore and the law of crack propagation in the marlite. By microscope and the references data, it can be concluded that the karst erosion process can be divided into rock minerals damage and rock structural damage. And the percent of karst erosion volume is named damage factor, which can be used to describe the quantify karst erosion degree of marlite. Through test, the rock mechanical properties in the different period of karst erosion are studied. Based on the damage mechanics theory and the test result, the relation between the karst erosion degree of marlite and weakening degree of mechanical properties is summarized. By numerical simulations, the karst erosive rock mass mechanics is verified. The conclusion is drawn as below: to the rock mass of marlite, the karst erosion damage made mechanics parameters variation, the deformation modulus, cohesion, and inter friction angle reduce as the negative exponent with the increasing of the karst erosion volume, however, the Poisson ratio increases as the positive exponent with the karst erosion volume increasing. It should be noticed that the deduced formulations are limited to the test data and certain conditions. It is suitable to the rock mass parametric weakening process after the karst erosion of marlite in Three Gorges Reservoir area. Based on the failure types of marlite slope in the field, the karst erosion and weathering process of rock mass are analyzed. And the evolution law of deformation and failure of the marlite mass is studied. The main failure feature of the marlite slope is the karst erosive structure subsidence mode in Three Gorges Reservoir area. The karst erosive structure subsidence mode is explained as follows: the rock mass undergoes the synthetic influence, such as weathering, unloading, corrosion, and so on, many pores and cavities have been formed in the rock mass interior, the rock mass quality is worsen and the rock mass structure is changed, and then the inherent structure of rock mass is collapsed under its gravity, therefore, the failure mode of compaction and subsidence take place. Finally, two examples are used to verify the rock mass parameters in Three Gorges Reservoir area, and the relationship between the marlite slope stability and the time of karst erosion is proposed.

内蒙古草地植被群落生物量的梯度特征研究

Relationship between biology and environment is always the theme of ecology. Transect is becoming one of the important methods in studies on relationship between global change and terrestrial ecosystems, especially for analysis of its driving factors. Inner Mongolia Grassland is the most important in China Grassland Transect brought forward by Yu GR. In this study, changes in grassland community biomass along gradients of weather conditions in Inner Mongolia was researched by the method of transect. Methods of regression about biomass were also compared. The transect was set from Eerguna county to Alashan county (38° 07' 35" ~50° 12' 20" N, 101° 55' 25" -120° 20' 46" E) in Inner Mongolia, China. The sample sites were mainly chosen along the gradient of grassland type, meadow steppe-* typical steppe-*desert steppe-*steppification desert-^desert. The study was carried out when grassland community biomass got the peak in August or September, 2003 and 2004. And data of 49 sample sites was gotten, which included biomass, mean annual temperature, annual precipitation, accumulated temperature above zero, annual hours of sunshine and other statistical and descriptive data. The aboveground biomass was harvested, and the belowground biomass was obtained by coring (30 cm deep). Then all the biomass samples were dried within (80 + 5) °C in oven and weighted. The conclusion is as follows: 1) From the northeast to the southwest in Inner Mongolia, along the gradient of grassland type, meadow steppe-*typical steppe-*desert steppe-*steppification desert-* desert, the cover degree of vegetation community reduces. 2) By unitary regression analysis, biomass is negatively correlated with mean annual temperature, s^CTC accumulated temperature, ^10°C accumulated temperature and annual hours of sunshine, among which mean annual temperature is crucial, and positively with mean annual precipitation and mean annual relative humidity, and the correlation coefficient between biomass and mean annual relative humidity is higher. Altitude doesn't act on it evidently. Result of multiple regression analysis indicates that as the primary restrictive factor, precipitation affects biomass through complicated way on large scale, and its impaction is certainly important. Along the gradient of grassland type, total biomass reduces. The proportion of aboveground biomass to total biomass reduces and then increases after desert steppe. The trend of below ground biomass's proportion to total biomass is adverse to that of aboveground biomass. 3) Precipitation is not always the only driving factor along the transect for below-/aboveground biomass ratio of different vegetation type composed by different species, and distribution of temperature and precipitation is more important, which is much different among climatic regions, so that the trend of below-/aboveground biomass ratio along the grassland transect may change much through the circumscription of semiarid region and arid region. 4) Among reproductive allocation of aboveground biomass, only the proportion of stem in total biomass notably correlates to the given parameters. Stem/leaf biomass ratio decreases when longitude and latitude increase, caloric variables decrease, and variables about water increase from desert to meadow steppe. The change trends are good modeled by logarithm or binomial equations. 5) 0'-10 cm belowground biomass highly correlates to environmental parameters, whose proportion to total biomass changes most distinctly and increases along the gradient from the west to the east. The deeper belowground biomass responses to the environmental change on the adverse trend but not so sensitively as the surface layer. Because the change value of 0~10 cm belowground biomass is always more than that of below 10 cm along the gradient, the deference between them is balanced by aboveground biomass's change by the resource allocation equilibrium hypothesis.

A Radial Basis Function Approach to Financial Time Series Analysis

Nonlinear multivariate statistical techniques on fast computers offer the potential to capture more of the dynamics of the high dimensional, noisy systems underlying financial markets than traditional models, while making fewer restrictive assumptions. This thesis presents a collection of practical techniques to address important estimation and confidence issues for Radial Basis Function networks arising from such a data driven approach, including efficient methods for parameter estimation and pruning, a pointwise prediction error estimator, and a methodology for controlling the "data mining'' problem. Novel applications in the finance area are described, including customized, adaptive option pricing and stock price prediction.

The Definition and Implementation of a Computer Programming Language Based on Constraints

The constraint paradigm is a model of computation in which values are deduced whenever possible, under the limitation that deductions be local in a certain sense. One may visualize a constraint 'program' as a network of devices connected by wires. Data values may flow along the wires, and computation is performed by the devices. A device computes using only locally available information (with a few exceptions), and places newly derived values on other, locally attached wires. In this way computed values are propagated. An advantage of the constraint paradigm (not unique to it) is that a single relationship can be used in more than one direction. The connections to a device are not labelled as inputs and outputs; a device will compute with whatever values are available, and produce as many new values as it can. General theorem provers are capable of such behavior, but tend to suffer from combinatorial explosion; it is not usually useful to derive all the possible consequences of a set of hypotheses. The constraint paradigm places a certain kind of limitation on the deduction process. The limitations imposed by the constraint paradigm are not the only one possible. It is argued, however, that they are restrictive enough to forestall combinatorial explosion in many interesting computational situations, yet permissive enough to allow useful computations in practical situations. Moreover, the paradigm is intuitive: It is easy to visualize the computational effects of these particular limitations, and the paradigm is a natural way of expressing programs for certain applications, in particular relationships arising in computer-aided design. A number of implementations of constraint-based programming languages are presented. A progression of ever more powerful languages is described, complete implementations are presented and design difficulties and alternatives are discussed. The goal approached, though not quite reached, is a complete programming system which will implicitly support the constraint paradigm to the same extent that LISP, say, supports automatic storage management.

Muscle function and body composition profile in adolescents withrestrictive anorexia nervosa: Does resistance training help?

The aim of this study was to analyze the effects of short-term resistance training on the body composition profile and muscle function in a group of Anorexia Nervosa restricting type (AN-R) patients. The sample consisted of AN-R female adolescents (12.8 ± 0.6 years) allocated into the control and intervention groups (n¼18 each). Body composition and relative strength were assessed at baseline, after 8 weeks and 4 weeks following the intervention. Body mass index (BMI) increased throughout the study (p = 0.011). Significant skeletal muscle mass (SMM) gains were found in the intervention group (p = 0.045, d = 0.6) that correlated to the change in BMI (r = 0.51, p < 0.031). Meanwhile, fat mass (FM) gains were significant in the control group (p = 0.047, d = 0.6) and correlated (r > 0.60) with change in BMI in both the groups. Significant relative strength increases (p < 0.001) were found in the intervention group and were sustained over time.

Fuzzy-Rough Sets Assisted Attribute Selection

R. Jensen and Q. Shen. Fuzzy-Rough Sets Assisted Attribute Selection. IEEE Transactions on Fuzzy Systems, vol. 15, no. 1, pp. 73-89, 2007.

Tolerance-based and Fuzzy-Rough Feature Selection.

R. Jensen and Q. Shen, 'Tolerance-based and Fuzzy-Rough Feature Selection,' Proceedings of the 16th International Conference on Fuzzy Systems (FUZZ-IEEE'07), pp. 877-882, 2007.