基于查询的新闻多文档自动摘要技术研究

针对新闻文本领域,该文提出一种基于查询的自动文本摘要技术,更加有针对性地满足用户信息需求。根据句子的TF-IDF、与查询句的相似度等要素,计算句子权重,并根据句子指示的时间给定不同的时序权重系数,使得最近发生的新闻内容具有更高的权重,最后使用最大边界相关的方法选择摘要句。通过与基于TF-IDF、Text-Rank、LDA等六种方法的对比,该摘要方法ROUGE评测指标上优于其他方法。从结合评测结果及摘要示例可以看出,该文提出的方法可以有效地从新闻文档集中摘取核心信息,满足用户查询内容的信息需求。     

基于特征选择和深度信念网络的文本情感分类算法

由于人类语言的复杂性，文本情感分类算法大多都存在因为冗余而造成的词汇量过大的问题。深度信念网络（DBN）通过学习输入语料中的有用信息以及它的几个隐藏层来解决这个问题。然而对于大型应用程序来说，DBN是一个耗时且计算代价昂贵的算法。针对这个问题，提出了一种半监督的情感分类算法，即基于特征选择和深度信念网络的文本情感分类算法（FSDBN）。首先使用特征选择方法（文档频率（DF）、信息增益（IG）、卡方统计（CHI）、互信息（MI））过滤掉一些不相关的特征从而使词汇表的复杂性降低；然后将特征选择的结果输入到DBN中，使得DBN的学习阶段更加高效。将所提算法应用到中文以及维吾尔语中，实验结果表明在酒店评论数据集上，FSDBN在准确率方面比DBN提高了1.6%，在训练时间上比DBN缩短一半。     

Reliability-based fault analysis models with industrial applications: A systematic literature review

Effective and early fault detection and diagnosis techniques have tremendously enhanced over the years to ensure continuous operations of contemporary complex systems, control cost, and enhance safety in assets-intensive industries, including oil and gas, process, and power generation. The objective of this work is to understand the development of different fault detection and diagnosis methods, their applications, and benefits to the industry. This paper presents a contemporary state-of-the-art systematic literature survey focusing on a comprehensive review of the models for fault detection and their industrial applications. This study uses advanced tools from bibliometric analysis to systematically analyze over 500 peer-reviewed articles on focus areas published since 2010. We first present an exploratory analysis and identify the influential contributions to the field, authors, and countries, among other key indicators.  A network analysis is presented to unveil and visualize the clusters of the distinguishable areas using a co-citation network analysis. Later, a detailed content analysis of the top-100 most-cited papers is carried out to understand the progression of fault detection and artificial intelligence–based algorithms in different industrial applications. The findings of this paper allow us to comprehend the development of reliability-based fault analysis techniques over time, and the use of smart algorithms and their success. This work helps to make a unique contribution toward revealing the future avenues and setting up a prospective research road map for asset-intensive industry, researchers, and policymakers.     

Hybrid Segmentation Scheme for Skin Features Extraction Using Dermoscopy Images

Objective and quantitative assessment of skin conditions is essential for cosmeceutical studies and research on skin aging and skin regeneration. Various handcraft-based image processing methods have been proposed to evaluate skin conditions objectively, but they have unavoidable disadvantages when used to analyze skin features accurately. This study proposes a hybrid segmentation scheme consisting of Deeplab v3+ with an Inception-ResNet-v2 backbone, LightGBM, and morphological processing (MP) to overcome the shortcomings of handcraft-based approaches. First, we apply Deeplab v3+ with an Inception-ResNet-v2 backbone for pixel segmentation of skin wrinkles and cells. Then, LightGBM and MP are used to enhance the pixel segmentation quality. Finally, we determine several skin features based on the results of wrinkle and cell segmentation. Our proposed segmentation scheme achieved a mean accuracy of 0.854, mean of intersection over union of 0.749, and mean boundary F1 score of 0.852, which achieved 1.1%, 6.7%, and 14.8% improvement over the panoptic-based semantic segmentation method, respectively.     

The effects of porosity,asphalt content and fiberglass incorporation on the tensile strength and resilient modulus of asphalt concrete blends

Asphalt concrete is the prevailing material used for road surface construction. Its adequate characteristics in providing stability, durability and driving safety are controlled by complex interactions between its components. Thus, it is important to estimate the sensitivity of asphalt concrete mechanical properties as a function of its volumetrics. For this study, different combinations between asphalt content (3.5, 5 and 7.5%) and porosity values (above 4%) were used in order to disassociate these properties. The influence of mixing in fiberglass (0.5%) was also analyzed. It was found that porosity is significantly more relevant than the asphalt content in the prediction of tensile strength and resilient modulus of fiber-free asphalt concretes. In fiber-reinforced mixtures, the mechanical properties are improved by increasing the asphalt content, which suggests a better bonding between fibers and aggregates. For both cases, decreasing porosity is beneficial. By grouping both sets of results, it was possible to create a unique theoretical curve for both the tensile strength (q_t) and the resilient modulus (RM). The RM/q_t ratio was 5800 for the fiber-free group, and 3900 for the fiber-reinforced group - suggesting a better fatigue life indicator for asphalt concretes when fibers are added.     

Antioxidant and antimicrobial potential of cajazeira leaves (Spondias mombin) extracts

Cajazeira leaves (Spondias mombin) have their highlighted use as antioxidant and natural antimicrobial, which justifies the objective of this work to evaluate the biological activities of different extracts. In order to evaluate the antioxidant and antimicrobial potentials of the cajazeira leaves, extractions at low pressure and high pressure were performed. The low pressure extractions (PLE) were carried out using Soxhlet (SOX) and tip ultrasound, using different solvents. Supercritical fluid extraction (SFE) was evaluated at temperature of 40?60°C and pressure of 150?300 bar besides extraction with cosolvent. Higher yields were obtained with the use of more polar solvents at LPE. The extracts obtained by SOX with ethanol and others polar solvents presented the best TPC values and antioxidant activity. The extracts at LPE with hexane and ethyl acetate and SFE presented better antimicrobial activity. Through liquid chromatography of high efficiency, it was possible to identify compounds with recognized biological activity, like ellagic acid, gallic acid and catechin.     

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。     

基于深度学习的文本自动摘要方案

针对自然语言处理（NLP）生成式自动摘要领域的语义理解不充分、摘要语句不通顺和摘要准确度不够高的问题，提出了一种新的生成式自动摘要解决方案，包括一种改进的词向量生成技术和一个生成式自动摘要模型。改进的词向量生成技术以Skip-Gram方法生成的词向量为基础，结合摘要的特点，引入词性、词频和逆文本频率三个词特征，有效地提高了词语的理解；而提出的Bi-MulRnn+生成式自动摘要模型以序列映射（seq2seq）与自编码器结构为基础，引入注意力机制、门控循环单元（GRU）结构、双向循环神经网络（BiRnn）、多层循环神经网络（MultiRnn）和集束搜索，提高了生成式摘要准确性与语句流畅度。基于大规模中文短文本摘要（LCSTS）数据集的实验结果表明，该方案能够有效地解决短文本生成式摘要问题，并在Rouge标准评价体系中表现良好，提高了摘要准确性与语句流畅度。     

Building a text collection for Urdu information retrieval

Urdu is a widely spoken language in the Indian subcontinent with over 300 million speakers worldwide. However, linguistic advancements in Urdu are rare compared to those in other European and Asian languages. Therefore, by following Text Retrieval Conference standards, we attempted to construct an extensive text collection of 85 304 documents from diverse categories covering over 52 topics with relevance judgment sets at 100 pool depth. We also present several applications to demonstrate the effectiveness of our collection. Although this collection is primarily intended for text retrieval, it can also be used for named entity recognition, text summarization, and other linguistic applications with suitable modifications. Ours is the most extensive existing collection for the Urdu language, and it will be freely available for future research and academic education.