Automated prediction system of household energy consumption in cities using web crawler and optimized artificial intelligence

The supply of electrical energy is critical to convenient and comfortable living. However, people consume a large amount of energy, contributing to an energy crisis and global warming, and damaging some ecological cycles. Residential electricity consumption has greater elasticity than industrial and business consumption; it therefore has high energy-saving potential. This work establishes an automated platform, which provides information about residential electricity consumption in each city in Taiwan. Machine learning was used to forecast future residential electricity demand. A nature-inspired optimization method was applied to enhance the accuracy of the best machine learner, yielding an even better hybrid ensemble model. Performance measures indicate that the resulting model is accurate and provides effective information for reference. An automatic web-based system based on the model was combined with a web crawler and scheduled to run automatically to provide information on monthly residential electricity consumption in each county and city. By providing energy consumption information across the country, power providers and government can discuss policy and set different goals for energy use. The results of this study can facilitate the early implementation of energy-saving and carbon emission-reducing in cities and aid utility companies in establishing energy conservation guidelines.     

Additional deepwater sampling of prey fish in Lake Michigan annual bottom trawl survey reveals new insights for depth distribution dynamics

Beginning in 2013, sites at the 128-m bottom depth contour were added to the sampling design of the annual Lake Michigan bottom trawl survey for prey fish, which has been conducted by the U.S. Geological Survey Great Lakes Science Center (GLSC) each fall since 1973, to better assess fish depth distributions in a changing ecosystem. The standard sampling design included bottom depths from 9 to 110 m, although the GLSC also sporadically conducted bottom trawl tows at the 128-m bottom depth contour during 1973–1988. Enactment of this new sampling design in 2013 revealed that mean biomass density of deepwater sculpins (Myoxocephalus thompsonii) at the 128-m depth exceeded the sum of mean biomass densities at shallower depths, indicating that the bulk of the deepwater sculpin population is residing in waters deeper than 110 m. Thus, our findings supported the hypothesis that the depth distribution of the deepwater sculpin population had shifted to deeper waters beginning in 2007, thereby explaining, at least in part, the marked decline in deepwater sculpin abundance since 2006 based on the standard sampling design. In contrast, our results did not support the hypothesis that the slimy sculpin (Cottus cognatus) population had shifted to deeper waters sometime after 2000. A portion of the burbot (Lota lota) population may have also shifted in depth distribution to waters deeper than 110 m after 2007, based on our results. Our findings have served as an impetus to further expand the range of depths sampled in our bottom trawl survey.     

基于卷积神经网络的单目深度估计

针对利用深度学习方法对街道图像进行深度估计，提出采用语义分割的方法解决深度图出现边界模糊等问题；估计深度通过左右视角图生成视差图进行无监督的训练。在网络模型中添加语义分割层，采取多个空洞卷积并行的结构增加感受野，同时减少了图像下采样的次数，降低了由于下采样带来的信息损失，使得的结果更加准确。这也是在深度估计中首次与空洞卷积相结合增加准确率。通过对KITTI街道数据集进行训练，与现有结果相比，除了增加检测准确性，降低错误率之外，使得效果图中的物体更加清晰，并且在效果图中还保留了一些原模型中被忽视掉的细节信息，将原始图像更加完整的表现出来。     

面向连续查询的敏感语义位置隐私保护方案

针对连续查询位置服务中构造匿名区域未考虑语义位置信息导致敏感隐私泄露问题，通过设计[(K,θ)]-隐私模型，提出一种路网环境下面向连续查询的敏感语义位置隐私保护方案。该方案利用Voronoi图将城市路网预先划分为独立的Voronoi单元，依据用户的移动路径和移动速度，选择具有相似特性的其他[K-1]个用户，构建匿名用户集；利用匿名用户集用户设定的敏感语义位置类型和语义安全阈值，以及用户所处语义位置的Voronoi单元，构建满足[(K,θ)]-隐私模型的语义安全匿名区域，可以同时防止连续查询追踪攻击和语义推断攻击。实验结果表明，与SCPA算法相比，该方案在隐私保护程度上提升约15%，系统开销上降低约20%。     

R2D2: A scalable deep learning toolkit for medical imaging segmentation

Deep learning has gained a significant popularity in recent years thanks to its tremendous success across a wide range of relevant fields of applications, including medical image analysis domain in particular. Although convolutional neural networks (CNNs) based medical applications have been providing powerful solutions and revolutionizing medicine, efficiently training of CNNs models is a tedious and challenging task. It is a computationally intensive process taking long time and rare system resources, which represents a significant hindrance to scientific research progress. In order to address this challenge, we propose in this article, R2D2, a scalable intuitive deep learning toolkit for medical imaging semantic segmentation. To the best of our knowledge, the present work is the first that aims to tackle this issue by offering a novel distributed versions of two well-known and widely used CNN segmentation architectures [ie, fully convolutional network (FCN) and U-Net]. We introduce the design and the core building blocks of R2D2. We further present and analyze its experimental evaluation results on two different concrete medical imaging segmentation use cases. R2D2 achieves up to 17.5× and 10.4× speedup than single-node based training of U-Net and FCN, respectively, with a negligible, though still unexpected segmentation accuracy loss. R2D2 offers not only an empirical evidence and investigates in-depth the latest published works but also it facilitates and significantly reduces the effort required by researchers to quickly prototype and easily discover cutting-edge CNN configurations and architectures.     

伪标签置信选择的半监督集成学习视频语义检测

在视频语义检测中，有标记样本不足会严重影响检测的性能，而且伪标签样本中的噪声也会导致集成学习基分类器性能提升不足。为此，提出一种伪标签置信选择的半监督集成学习算法。首先，在三个不同的特征空间上训练出三个基分类器，得到基分类器的标签矢量；然后，引入加权融合样本所属某个类别的最大概率与次大概率的误差和样本所属某个类别的最大概率与样本所属其他各类别的平均概率的误差，作为基分类器的标签置信度，并融合标签矢量和标签置信度得到样本的伪标签和集成置信度；接着，选择集成置信度高的样本加入到有标签的样本集，迭代训练基分类器；最后，采用训练好的基分类器集成协作检测视频语义概念。该算法在实验数据集UCF11上的平均准确率到达了83.48%，与Co-KNN-SVM算法相比，平均准确率提高了3.48个百分点。该算法选择的伪标签能体现样本所属类别与其他类别的总体差异性，又能体现所属类别的唯一性，可减少利用伪标签样本的风险，有效提高视频语义概念检测的准确率。     

Title加TextRank抽取关键句的情感分类研究

考虑到不同句子对判断文档情感倾向的重要程度不同，因而区分文档的关键句和细节句将有助于提高情感分类的性能。同时，考虑到Title和上下文信息，提出了一种基于Title和加权TextRank抽取关键句的情感分析方法SKTT，实现了高效的情感分析。根据文档Title的情感权重计算Title贡献度，考虑到标点和语义规则对情感倾向的影响；根据加权TextRank算法思想，在文档正文中构建了一个情感句有向图来提取关键句；计算所有关键句的情感倾向进行情感分类。在4个领域上进行实验，实验结果表明，该SKTT方法性能明显优于Baseline，具有高效性。     

物联网环境下基于情境的语义Web服务选择

当前Web服务海量增加，物联网应用技术快速发展、不断普及，而现有的Web服务选择算法低效、用户匹配度低。针对该问题提出一种物联网环境下基于情境的语义Web服务选择方法。该方法应用QoS参数的无量纲化与语义Web服务动态选择方式，将物联网环境下服务与语义Web服务相结合，并根据用户需求针对QoS选择最优的服务集。实验表明，该方法能有效地提高用户服务动态选择的成功率。     

基于文本与语义相关性分析的图像检索

为了更加有效地检索到符合用户复杂语义需求的图像，提出一种基于文本描述与语义相关性分析的图像检索算法。该方法将图像检索分为两步：基于文本语义相关性分析的图像检索和基于SIFT特征的相似图像扩展检索。根据自然语言处理技术分析得到用户文本需求中的关键词及其语义关联，在选定图像库中通过语义相关性分析得到“种子”图像；接下来在图像扩展检索中，采用基于SIFT特征的相似图像检索，利用之前得到的“种子”图像作为查询条件，在网络图像库中进行扩展检索，并在结果集上根据两次检索的图像相似度进行排序输出，最终得到更加丰富有效的图像检索结果。为了证明算法的有效性，在标准数据集Corel5K和网络数据集Deriantart8K上完成了多组实验，实验结果证明该方法能够得到较为精确地符合用户语义要求的图像检索结果，并且通过扩展算法可以得到更加丰富的检索结果。     

The effect of cognitive and affective aspects on usability

Studies on customer’s needs, desires and preferences have become highly important in the product design and development process. One consideration in usability is the cognitive aspect, which is related to the accommodation and evaluation of human cognitive capabilities, limitations and tendencies. In addition to the cognitive aspect, a recent study has shown that the affective aspect has been considered in the evaluation of product usability. Thus, both cognitive and affective aspects are deemed to be important for product design and the development process. Inherently, both aspects deliver complete human and product interaction and experience. However, studies that consider the affective process as a complement to the cognitive process for usability are relatively rare. To address this gap, this study discusses how an integrative framework of the cognitive and affective aspects can be applied to a product for usability assessment via empirical studies on e-commerce and e-learning platforms. The sample involved 230 respondents, using purposive sampling. The result shows that both cognitive and affective aspects have a significant effect, although with different weights. The affective aspect has been shown to improve product usability and user’s acceptance.