YOLSO: You Only Look Small Object

Small object detection is challenging and far from satisfactory. Most general object detectors suffer from two critical issues with small objects: (1) Feature extractor based on classification network cannot express the characteristics of small objects reasonably due to insufficient appearance information of targets and a large amount of background interference around them. (2) The detector requires a much higher location accuracy for small objects than for general objects. This paper proposes an effective and efficient small object detector YOLSO to address the above problems. For feature representation, we analyze the drawbacks in previous backbones and present a Half-Space Shortcut(HSSC) module to build a background-aware backbone. Furthermore, a coarse-to-fine Feature Pyramid Enhancement(FPE) module is introduced for layer-wise aggregation at a granular level to enhance the semantic discriminability. For loss function, we propose an exponential L1 loss to promote the convergence of regression, and a focal IOU loss to focus on prime samples with high classification confidence and high IOU. Both of them significantly improves the location accuracy of small objects. The proposed YOLSO sets state-of-the-art results on two typical small object datasets, MOCOD and VeDAI, at a speed of over 200 FPS. In the meantime, it also outperforms the baseline YOLOv3 by a wide margin on the common COCO dataset.     

Influence of temperature on acoustic emission source location accuracy in underground structure

An acoustic emission (AE) experiment was carried out to explore the AE location accuracy influenced by temperature. A hollow hemispherical specimen was used to simulate common underground structures. In the process of heating with the flame, the pulse signal of constant frequency was stimulated as an AE source. Then AE signals received by each sensor were collected and used for comparing localization accuracy at different temperatures. Results show that location errors of AE keep the same phenomenon in the early and middle heating stages. In the later stage of heating, location errors of AE increase sharply due to the appearance of cracks. This provides some beneficial suggestions on decreasing location errors of structural cracks caused by temperature and improves the ability of underground structure disaster prevention and control.     

基于Bert预训练模型的虚假新闻文本检测

虚假新闻剥夺了人们获取真相的权利,也给社会和国家带来了许多危害。文中以虚假新闻文本为例,分析和验证了多种预训练语言模型在虚假新闻文本分类上的分类效果。经实验证明,相较于其他语言模型,Bert预训练语言模型取得了较好的结果,预测准确率为86.97%,召回率为88.12%,F1值为87.54%。     

基于声源定位的高校在线巡课系统设计与研究

随着网络建设以及信息化教学方法在高校教学过程中的应用普及,越来越多的高校使用在线巡课系统对教师的教学过程进行跟踪和管理,以便发现课堂教学中的亮点、问题和不足。文章提出的在线巡课系统,基于声源定位的技术,对传统的在线巡课系统做出了改进,解决了已有巡课系统中“只闻其声,不见其人”的问题,能够更直观地跟踪到教师的教学过程,包括教学行为以及师生互动过程,有效提升教务人员巡课效果和体验感受。     

A hybrid knowledge and ensemble classification approach for prediction of venous thromboembolism

Clinical narratives such as progress summaries, lab reports, surgical reports, and other narrative texts contain key biomarkers about a patient's health. Evidence-based preventive medicine needs accurate semantic and sentiment analysis to extract and classify medical features as the input to appropriate machine learning classifiers. However, the traditional approach of using single classifiers is limited by the need for dimensionality reduction techniques, statistical feature correlation, a faster learning rate, and the lack of consideration of the semantic relations among features. Hence, extracting semantic and sentiment-based features from clinical text and combining multiple classifiers to create an ensemble intelligent system overcomes many limitations and provides a more robust prediction outcome. The selection of an appropriate approach and its interparameter dependency becomes key for the success of the ensemble method. This paper proposes a hybrid knowledge and ensemble learning framework for prediction of venous thromboembolism (VTE) diagnosis consisting of the following components: a VTE ontology, semantic extraction and sentiment assessment of risk factor framework, and an ensemble classifier. Therefore, a component-based analysis approach was adopted for evaluation using a data set of 250 clinical narratives where knowledge and ensemble achieved the following results with and without semantic extraction and sentiment assessment of risk factor, respectively: a precision of 81.8% and 62.9%, a recall of 81.8% and 57.6%, an F measure of 81.8% and 53.8%, and a receiving operating characteristic of 80.1% and 58.5% in identifying cases of VTE.     

基于微博用户兴趣话题的相似用户挖掘

相似用户挖掘是提高社交网络服务质量的重要途径，在面向大数据的社交网络时代，准确的相似用户挖掘对于用户和互联网企业等都有重要的意义，而根据用户自己的兴趣话题挖掘的相似用户更符合相似用户的要求。提出了一种基于用户兴趣话题进行相似用户挖掘的方法。该方法首先使用TextRank话题提取方法对用户进行兴趣话题提取，再对用户发表内容进行训练，计算出所有词之间的相似度。提出CP（Corresponding Position similarity）、CPW（Corresponding Position Weighted similarity）、AP（All Position similarity）、APW（All Position Weighted similarity）四种用户兴趣话题词相似度计算方法，通过用户和相似用户间关注、粉丝重合率验证相似用户挖掘效果，APW similarity的相似用户的关注/粉丝重合百分比为1.687%，优于提出的其他三种算法，分别提高了26.3%、2.8%、12.4%，并且比传统的文本相似度方法Jaccard相似度、编辑距离算法、余弦相似度分别提高了20.4%、21.2%、45.0%。因此APW方法可以更加有效地挖掘出用户的相似用户。     

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。     

The generalized discrete ‐centroid problem

The ‐centroid problem or leader–follower problem is generalized considering different customer choice rules where a customer may use facilities belonging to different firms, if the difference in travel distance (or time) is small enough. Assuming essential goods, some particular customer choice rules are analyzed. Linear programming formulations for the generalized ‐medianoid and ‐centroid problems are presented and an exact solution approach is applied. Some computational examples are included.     

基于多元神经网络融合的分布式资源空间文本分类研究

针对实体产业对科技资源的服务需求,以服务效应作为资源文本分类标准,提出一种基于多元神经网络融合的分布式资源空间文本分类模型。设计了包含词嵌入层、卷积层、双向门控循环单元层、注意力机制层和softmax层的多元神经网络通路;在此基础上采用基于需求—效应—资源分类策略,完成了从定性科技资源需求到定量资源服务效应求解,再到定性科技资源输出的映射变换,重点解决了分布式科技资源局部和全局语义特征形式多样、文本长距离依赖特征显著、重要资源信息难以准确识别的问题,进而从分布式科技资源空间中快速准确地获取效应知识,提升实体产业产品研发效率和创新能力;通过万方专利科技资源数据集验证了所提方法的可行性和有效性,为更加全面地挖掘资源文本特征和按需服务实体产业提供了一种新的思路和手段。     

Restructuring bank networks after mergers and acquisitions: A capacitated delocation model for closing and resizing branches

During restructuring processes, due to mergers and acquisitions, banks frequently face the problem of having redundant branches competing in the same market. In this work, we introduce a new Capacitated Branch Restructuring Model which extends the available literature in delocation models. It considers both closing down and long term operations׳ costs, and addresses the problem of resizing open branches in order to maintain a constant service level. We consider, as well, the presence of competitors and allow for ceding market share whenever the restructuring costs are prohibitively expensive.We test our model in a real life scenario, obtaining a reduction of about 40% of the network size, and annual savings over 45% in operation costs from the second year on. We finally perform a sensitivity analysis on critical parameters. This analysis shows that the final design of the network depends on certain strategic decisions concerning the redundancy of the branches, as well as their proximity to the demand nodes and to the competitor׳s branches. At the same time, this design is quite robust to changes in the parameters associated with the adjustments on service capacity and with the market reaction.