基于推理注意力机制的二阶段网络图像修复

现有的图像修复方法在处理大面积缺失或高度纹理化的图像时，通常会产生扭曲的结构或与周围区域不一致的模糊纹理，无法重建合理的图像结构。为此，提出了一种基于推理注意力机制的二阶段网络图像修复方法。首先通过边缘生成网络生成合理的幻觉边缘信息，然后在图像补全网络完成图像的重建工作。为了进一步生成视觉效果更逼真的图像，提高图像修复的精确度，在图像补全网络采用推理注意力机制，有效控制了生成特征的不一致性，从而生成更有效的信息。所提方法在多个数据集上进行了实验验证，结果表明该图像修复方法的结构相似性指数达到了88.9%，峰值信噪比达到了25.56 dB，与现有的图像修复方法相比，该方法具有更高的图像修复精确度，生成的图像更逼真。     

基于ANSYS Workbench的天然气渐扩管冲蚀磨损仿真模拟

为了研究含砂天然气集输管道渐扩管的冲蚀磨损现象。运用ANSYS软件建立渐扩管流场的Euler多相流、流固耦合及冲蚀模型，分析渐扩管的应力分布、流动特性及冲蚀效果，并与突扩管进行比较；通过改变管道入口流速、颗粒粒径、颗粒质量流率等参数，对渐扩管流场特性及冲蚀规律进行分析。结果表明，渐扩管应力集中在喉部及出口管段内壁面，冲蚀区域集中在出口管段下表面及出口处，出口管段冲蚀区域呈“椭圆形”，出口处冲蚀区域呈“U”形；突扩管应力集中在进口管段与出口管段交界处的上壁面，冲蚀区域位于进口管段；渐扩管冲蚀速率随入口流速增加先增大后减小再增大，随颗粒粒径的增加先增大后减小最终趋于稳定，随颗粒质量流率增加呈线性增长；相同条件下，突扩管最大等效应力是渐扩管的1.73倍，变形位移是渐扩管的1.77倍，而渐扩管冲蚀较为严重，最大冲蚀速率是突扩管的1.7倍。     

基于改进模糊有色Petri网的综合能源系统能耗动态过程模型及应用

针对综合能源系统在供能过程中的能耗动态特点,通过改进的Petri网方法,根据特定的建模规则,建立了一种基于改进模糊有色Petri网的综合能源系统能耗动态过程模型,并在Stateflow平台上利用实际算例进行了仿真分析。结果表明,所建模型合理、有效,有一定的分析计算特性和图形化表现能力,为进一步对系统需求侧规划和多能设备的合理匹配提供了理论基础和技术支持。     

借鉴逻辑学方法提高临床思维能力

目的：正确的临床决策必须遵循逻辑规律。本文旨在剖析用逻辑学方法提高临床思维训练的可行性。方法：进行临床医学和逻辑学的跨学科整合研究，重点探讨三类常用逻辑方法：演绎法、归纳法和溯因推理法在临床实践中的应用。 结果：发现演绎、溯因和归纳推理分别对应于：①评估临床资料；②提出诊断假设和③完善初步诊断这三种临床思维活动。发现并概括临床思维和医疗实践中包含的逻辑规律，有助于提高思维的准确性、有效性和清晰性。结论：掌握临床思维的逻辑规律，并在中针对性地加以应用，有助于培养医生的思维和决策能力。     

基于改进的Random Subspace的客户投诉分类方法

电信业的客户投诉不断增多而又亟待高效处理。针对电信客户投诉数据的特点，提出了一种面向高维数据的改进的集成学习分类方法。该方法综合考虑客户投诉中的文本信息及客户通讯状态信息，基于Random Subspace方法，以支持向量机（Support Vector Machine，SVM）为基分类器，采用证据推理（Evidential Reasoning，ER）规则为一种新的集成策略，构造分类模型对电信客户投诉进行分类。所提模型和方法在某电信公司客户投诉数据上进行了验证，实验结果显示该方法能够显著提高客户投诉分类的准确率和投诉处理效率。     

Effects of feeding different amounts of milk replacer on growth performance and nutrient digestibility in Holstein calves to 2 months of age using different weaning strategies

Growth and the digestibility of nutrients can be greatly affected by diet preweaning and the rate at which calves are weaned. A 2 × 2 factorial design [moderate (MOD) or high (HI) milk replacer (MR) feeding rates and abrupt (AB) or gradual (GR) weaning] was used to compare these effects. Calves (n = 50) were randomly assigned to 1 of 4 treatments: MOD-AB, MOD-GR, HI-AB, and HI-GR. Calves assigned to MOD-AB were fed 0.66 kg of MR for the first 42 d and then 0.33 kg for the last 7 d; those assigned to MOD-GR were fed 0.66 kg of MR for 28 d, 0.33 kg for 14 d, and 0.17 kg for the last 7 d; those assigned to HI-AB were fed 0.66 kg of MR for 7 d, 0.82 kg for 7 d, 1.1 kg for 28 d, and 0.66 kg for the last 7 d; and those assigned to HI-GR were fed 0.66 kg of MR for 7 d, 0.82 kg for 7 d, 1.1 kg for 14 d, 0.66 kg for 14 d, and 0.33 kg for the last 7 d. All calves received the same MR [25% crude protein (CP), 17% fat; dry matter (DM) basis] and were given ad libitum access to water and a textured starter (42% starch and 20% CP). On d 26 to 30 and d 45 to 49, a fecal sample was taken from 5 calves in each treatment via the rectum to estimate apparent digestibility coefficients (dC). Apparent dC of DM, organic matter, and fat were greater for HI versus MOD calves. Apparent dC of neutral detergent fiber, acid detergent fiber, and sugar were greater for MOD versus HI calves. Apparent dC of DM was greater for AB versus GR calves [90.9, 89.0 ± 0.5384 (standard error)], and the apparent dC of acid detergent fiber was greater for GR versus AB calves (39.5, 32.3 ± 1.67). Feed efficiency was greater for HI versus MOD and AB versus GR. There were no significant differences between CP or starch dC based on treatment, and no interactions were observed. Starter consumption was greater for MOD calves compared with HI calves, and GR calves consumed more than AB calves. These results suggest that providing calves high amounts of MR preweaning enhanced readily available nutrient dC, but providing moderate amounts of MR resulted in increased fibrous fraction dC. Accordingly, the AB weaning strategy had higher dC for DM and organic matter, but there was a depression in fiber dC.     

Short term stock selection with case-based reasoning technique

Stock selection is an important decision making problem. Trading strategies and rules based on fundamental and technical analysis can be used for decision making process. In this paper, we propose an intelligent stock selection method, which is called case-based reasoning (CBR). This technique uses the fundamental and technical indicators to identify the winning stocks around the earning announcements. CBR method is compared with other artificial intelligence techniques such as multi layer perceptron (MLP), decision trees (QUEST, Classification and Regression Trees, C5), generalized rule induction (GRI) and logistic regression. We show that the performance of CBR is better than the performance of other techniques in terms of classification accuracy, average return, Sharpe ratio and ideal profit.     

一种基于证据推理的视频运动目标融合检测算法

为了解决复杂场景下传统的运动目标检测问题, 利用证据推理—–谨慎有序加权平均方法(COWA-ER), 提出一种综合使用混合高斯、均值滤波和码本的多方法融合的检测方法. 该融合检测算法以上述3 种检测方法为准则建立一个多准则决策框架, 通过双阈值检测法来表征检测过程中的不确定性, 最终利用COWA-ER 方法进行决策级融合, 实现多种方法的优势互补. 实验表明, 所提出的融合检测算法具有更理想的目标检测效果, 能有效应对诸如阴影及光照突变等问题对检测性能的影响.