Locally Weighted Learning

This paper surveys locally weighted learning, a form of lazy learning and memory-based learning, and focuses on locally weighted linear regression. The survey discusses distance functions, smoothing parameters, weighting functions, local model structures, regularization of the estimates and bias, assessing predictions, handling noisy data and outliers, improving the quality of predictions by tuning fit parameters, interference between old and new data, implementing locally weighted learning efficiently, and applications of locally weighted learning. A companion paper surveys how locally weighted learning can be used in robot learning and control.     

小型RFCO2激光器的电路保护

本文详细分析研究了小型射频CO2激光器经常出现的故障现象(过压、负载短路故障),并介绍了相应的保护措施,对实际工作有显着指导意义。     

薄层反射波非零炮检距的属性特征

 本文从薄层顶、底界面反射波的走时入手，研究了薄层反射波在非零炮检距集上出现的干涉现象，导出了薄层反射波在出现干涉时的炮检距计算公式，说明薄层干涉出现的炮检距是多变量函数，即随着埋藏深度h、地层速度v、薄层厚度Δh以及地震波的主频f*的变化而改变。本文重点分析了地层厚度Δh和地震子波主频f*对反射波振幅特性和频率特性的影响。在充分分析薄层反射的振幅、频谱随炮检距变化关系的基础上，指出在应用叠前信息（如AVO分析）进行储层预测等方面的研究时，应充分考虑薄层反射波的干涉和调谐现象带来的振幅和频率特征的变化。     

宽线性共轭梯度算法及其在盲多用户检测中的应用

本文提出了一种宽线性共轭梯度算法，分析了算法在均值意义下的收敛性，并给出了算法稳定条件。仿真表明，新算法应用于码分多址系统多址干扰抑制时的性能优于Yin（2003）及Schober（2004）所提出的算法。     

三维地震勘探中强声波干扰的压制

在三维地震资料采集中产生的声波为非线性干扰波,能量强,频谱宽,利用常规的分频去噪和线性压噪方法难以将其去除,使后续的资料处理效果受到影响。基于声波具有非常稳定的传播速度这一特点,提出了一种利用非线性变换手段压制强声波干扰的方法。方法原理是,首先对三维地震资料中含声波到达时间的某段记录(时窗)进行转换,将非线性声波时距关系转换为速度趋于无穷、波数趋于零的一组噪声;然后应用f-k滤波对这组噪声进行压制;最后,再通过逆转换得到声波压制后的结果。该方法易于实现,并只对声波到达时间处的小范围数据进行处理,数据量小,计算效率高。对实际三维地震资料的处理结果表明,该方法压制强声波干扰的效果非常明显。     

用于微波/射频集成电路的一种新型低损耗介质——多孔硅及氧化多孔硅厚膜

提出在单片微波集成电路(MMIC)中用多孔硅/氧化多孔硅厚膜作微波无源器件的低损耗介质膜.研究了厚度为70μm的多孔硅/氧化多孔硅厚膜在低阻硅衬底上的形成,这层厚膜增加了衬底的电阻率,减少了微波的有效介质损耗.通过测量在低阻硅衬底上形成的氧化多孔硅厚膜上的共平面波导的微波特性,证明了在低阻硅衬底上用厚膜氧化多孔硅可以提高共平面传输线(CPW)的微波特性.     

Nonconscious control mimics a purposeful strategy: Strength of Stroop-like interference is automatically modulated by proportion of compatible trials.

The magnitude of the Stroop effect is known to be modulated by the proportion of trials on which the irrelevant word and relevant ink color correspond. This has often been attributed to a conscious strategy of increased (or decreased) reliance on the irrelevant words when these are more likely (or less likely) to correspond to the ink colors. However, the present data from a Stroop-like task involving successively presented arrows indicate instead that this type of modulation can be automatic because it can occur even if the irrelevant stimuli are not phenomenally visible. In this case participants cannot determine the proportion of compatible trials to direct their strategy. An automatic, item-specific associative interpretation can account for these findings. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The influence of directional associations on directed forgetting and interference.

Two experiments examined how cross-list directional associations influenced list-method directed forgetting and the degree of interference observed on each list. Each List 1 item had a (a) bidirectionally related item on List 2 (chip ←→ potato), (b) forward association with an item on List 2 (chip → wood), (c) backward association from an item on List 2 (chip ← chisel), or (d) no relationship with List 2 items. The results revealed that associative relationships that eliminated retroactive interference in the baseline condition also eliminated the directed forgetting costs. In contrast, associative relationships did not affect List 2 recall in the forget group, which remained unchanged across experimental conditions. However, certain conditions reduced proactive interference in the remember group, thereby eliminating the benefits of directed forgetting. The directed forgetting costs and benefits were observed independently of each other. The authors propose that these effects emerged from a combination of item and context strengthening induced by different associative directions. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Still No Evidence for Perfect Timesharing With Two Ideomotor-Compatible Tasks: A Reply to Greenwald (2003).

For 30 years, A. G. Greenwald and H. G. Shulman's (1973) psychological refractory period (PRP) study has been cited as evidence for perfect timesharing with ideomotor (IM)-compatible tasks. Recently, M.-C. Lien, R. W. Proctor, and P. A. Allen (2002) failed to replicate these results and concluded that IM compatibility is neither necessary nor sufficient to eliminate the PRP effect. A. G. Greenwald (2003) attributed Lien et al.'s nonreplication to the use of (a) a non-IM-compatible task, (b) varied trial spacing, and/or (c) inappropriate instructions. The authors of the present article argue that the first 2 factors are not critical and that instructions merely affect the criterion for speed versus accuracy. In each of Greenwald's experiments, dual-task costs were evident on response time or error rates. Furthermore, the small dual-task costs in his study are consistent with a bottleneck model. Thus, Greenwald (2003) does not provide evidence that IM-compatible tasks enable perfect timesharing. (PsycINFO Database Record (c) 2010 APA, all rights reserved)