强化学习中的时间差分法在用户学习方面的应用

随着Internet上信息量的与日俱增，用户查找自己感兴趣的信息变得越来越困难。文章提出了一种使用强化学习中的时间差算法来学习用户的新学习算法，并通过实验与其它学习方法作了比较，取得了明显的成效。     

一种基于EBD模型的多板系统信号完整性仿真方案

文章简述了当前仿真软件对基于EBD(ELECTRICALBOARDDESCRIPTION)模型的多板系统信号完整性问题仿真中存在的问题,并结合Cadence仿真软件提出了自己的一种仿真解决方案。与原有的方案相比,该文的方法在保证了适合前仿真与约束建立的前提下,更加全面准确地反映了模型信息,并结合实例给出了相应的仿真结果。     

低应变反射波法检测桩基结构完整性的应用

桩基检测是一项非常重要的工作，其不可预见影响因素非常复杂。例如，检测桩的桩头处理、传感器的安装、激振频率的选择情况以及桩周土阻力等，都对检测数据产生一定的影响，这就要求检测人员在检测过程中，分析克服外界不利影响因素，寻求最佳检测手段，追求最准确的判断。文章通过对低应变反射波法在桩基检测中的应用分析，针对现场数据采集及检测波形判读中一些值得注意的问题，作了初步的探讨。     

Watershed scale temporal and spatial stability of soil moisture and its role in validating satellite estimates

Watershed scale soil moisture estimates are necessary to validate current remote sensing products, such as those from the Advanced Microwave Scanning Radiometer (AMSR). Unfortunately, remote sensing technology does not currently resolve the land surface at a scale that is easily observed with ground measurements. One approach to validation is to use existing soil moisture measurement networks and scale these point observations up to the resolution of remote sensing footprints. As part of the Soil Moisture Experiment 2002 (SMEX02), one such soil moisture gaging system in the Walnut Creek Watershed, Iowa, provided robust estimates of the soil moisture average for a watershed throughout the summer of 2002. Twelve in situ soil moisture probes were installed across the watershed. These probes recorded soil moisture at a depth of 5 cm from June 29, 2002 to August 19, 2002. The sampling sites were analyzed for temporal and spatial stability by several measures including mean relative difference, Spearman rank, and correlation coefficient analysis. Representative point measurements were used to estimate the watershed scale (∼25 km) soil moisture average and shown to be accurate indicators with low variance and bias of the watershed scale soil moisture distribution. This work establishes the validity of this approach to provide watershed scale soil moisture estimates in this study region for the purposes of satellite validation with estimation errors as small as 3%. Also, the potential sources of error in this type of analysis are explored. This study is a first step in the implementation of large-scale soil moisture validation using existing networks such as the Soil Climate Analysis Network (SCAN) and several Agricultural Research Service watersheds as a basis for calibrating satellite soil moisture products, for networks design, and designing field experiments.     

宁夏农田土壤耕层养分含量的时空变化特征

根据土壤普查资料和实际采样测定结果,探讨了宁夏山区和灌区农田土壤耕层有机质和氮磷钾养分含量的时空变化特征。结果表明,在1970年、1985年和2000年3个时段内,土壤有机质、有效磷和碱解氮含量呈显著或较显著增长态势,土壤速效钾含量近15年在灌区明显下降,土壤氮磷钾全量养分在30年间没有明显变化。从两个地区比较来看,除土壤全磷量外,土壤有机质和其它养分含量都是灌区显著高于山区,而变异系数除土壤全钾外都是山区显著大于灌区。     

Surface integrity study of WEDM with various wire electrodes: experiments and analysis

Abstract

Wire electrical discharge machining (WEDM) is always significant for its high-precision machining. However, due to the generation of high discharge energy during machining, machined surfaces are often got distorted. These might be upgraded by choosing the correct tool with proper machining condition. The effects of the electrode materials and process parameters on different responses of WEDM like average surface roughness, recast layer thickness, and surface morphology are systematically examined here to enhance the knowledge of WEDM and its correlation with electrode property. The experiments have been carried out on one of the expensive steel namely Maraging steel 300 due to its applicability in tooling and aerospace industries. Plain brass wire, zinc-coated brass wire (ZCB), and silver-coated brass (SCB) wires are used as a tool electrode for analysis. Comparative experimental studies prove that among BW, ZCB, and SCB, the overall performance of SCB is commendable owing to the high-quality surface considering control parameters in low discharge energy level. However, the second-best performance is shown by ZCB.     

LWR Fuel Safety Research with Particular Emphasis on RIA/LOCA and Other Conditions

Fuel safety research at Japan Atomic Energy Research Institute (JAERI) is reviewed on the major subjects including studies on fuel behavior under postulated Reactivity Initiated Accident (RIA), postulated Loss of Coolant Accident (LOCA) and normal operating conditions. Nuclear Safety Research Reactor (NSRR) at JAERI has been utilized extensively for the studies of fuel behavior under RIA conditions. For the studies of fuel rod and cladding behavior under LOCA conditions, outpile experiments were conducted. The work on this subject has been concluded. Pellet Cladding Interaction (PCI) has been major subject on fuel integrity study during normal operating conditions. Irradiation experiments at Halden Boiling Water Reactor (HBWR) as well as code development are described.     

Alternating temporal dominance of sensations and liking scales during the intake of a full portion of an oral nutritional supplement

Measurement of liking with consumers is often performed on a single sip or mouthful of the product and results in a single liking score that does not necessarily reflect the liking over consumption of a whole product. A dynamic approach could provide this type of information. Thus, obtaining reasons why the liking score is changing over consumption could help to gain insights for product improvement. Recently, Thomas, Visalli, Cordelle, and Schlich (2015) suggested collecting with the same consumers liking scores dynamically and Temporal Dominance of Sensations (TDS) during mono-intakes of flavored cheeses, in two independent sessions. This new type of sensory data has been analysed at individual level to identify “temporal drivers of liking” for each product.The present paper further develops this method consisting of collecting TDS and dynamic liking data in the same session and during the consumption of the full portion of a product. The products used to experiment this new method were two commercial oral nutritional supplements (ONS) assessed by 65 consumers over two lab sessions separated by one week.The protocol consisted of a TDS evaluation of 5 flavor, 2 texture and 3 taste sensations during each sip, immediately followed by a liking scale. This protocol was repeated over sips. This method, named “Alternated Temporal Drivers of Liking” (A-TDL), required that consumers record liking changes and do TDS over sips during food consumption. Results showed that one of the two ONS was more liked than the other and consumed in a larger volume resulting in a higher energy intake. In contrast, the other product resulted in a stronger thirst and presented more negative temporal drivers of liking (dry, metallic, filming) over the consecutive sips.     

Temporal contexts: Effective text classification in evolving document collections

The management of a huge and growing amount of information available nowadays makes Automatic Document Classification (ADC), besides crucial, a very challenging task. Furthermore, the dynamics inherent to classification problems, mainly on the Web, make this task even more challenging. Despite this fact, the actual impact of such temporal evolution on ADC is still poorly understood in the literature. In this context, this work concerns to evaluate, characterize and exploit the temporal evolution to improve ADC techniques. As first contribution we highlight the proposal of a pragmatical methodology for evaluating the temporal evolution in ADC domains. Through this methodology, we can identify measurable factors associated to ADC models degradation over time. Going a step further, based on such analyzes, we propose effective and efficient strategies to make current techniques more robust to natural shifts over time. We present a strategy, named temporal context selection, for selecting portions of the training set that minimize those factors. Our second contribution consists of proposing a general algorithm, called Chronos, for determining such contexts. By instantiating Chronos, we are able to reduce uncertainty and improve the overall classification accuracy. Empirical evaluations of heuristic instantiations of the algorithm, named WindowsChronos and FilterChronos, on two real document collections demonstrate the usefulness of our proposal. Comparing them against state-of-the-art ADC algorithms shows that selecting temporal contexts allows improvements on the classification accuracy up to 10%. Finally, we highlight the applicability and the generality of our proposal in practice, pointing out this study as a promising research direction.