Leveraging Uncertainty for Depth-Aware Hierarchical Text Classification

Hierarchical Text Classification(HTC)aims to match text to hierarchical labels.Existing methods overlook two critical issues:first,some texts cannot be fully matched to leaf node labels and need to be classified to the correct parent node instead of treating leaf nodes as the final classification target.Second,error propagation occurs when a misclassification at a parent node propagates down the hierarchy,ultimately leading to inaccurate predictions at the leaf nodes.To address these limitations,we propose an uncertainty-guided HTC depth-aware model called DepthMatch.Specifically,we design an early stopping strategy with uncertainty to identify incomplete matching between text and labels,classifying them into the corresponding parent node labels.This approach allows us to dynamically determine the classification depth by leveraging evidence to quantify and accumulate uncertainty.Experimental results show that the proposed DepthMatch outperforms recent strong baselines on four commonly used public datasets:WOS(Web of Science),RCV1-V2(Reuters Corpus Volume I),AAPD(Arxiv Academic Paper Dataset),and BGC.Notably,on the BGC dataset,it improvesMicro-F1 andMacro-F1 scores by at least 1.09%and 1.74%,respectively.

New Solution Generation Strategy to Improve Brain Storm Optimization Algorithm for Classification

As a new intelligent optimization method,brain storm optimization(BSO)algorithm has been widely concerned for its advantages in solving classical optimization problems.Recently,an evolutionary classification optimization model based on BSO algorithm has been proposed,which proves its effectiveness in solving the classification problem.However,BSO algorithm also has defects.For example,large-scale datasets make the structure of the model complex,which affects its classification performance.In addition,in the process of optimization,the information of the dominant solution cannot be well preserved in BSO,which leads to its limitations in classification performance.Moreover,its generation strategy is inefficient in solving a variety of complex practical problems.Therefore,we briefly introduce the optimization model structure by feature selection.Besides,this paper retains the brainstorming process of BSO algorithm,and embeds the new generation strategy into BSO algorithm.Through the three generation methods of global optimal,local optimal and nearest neighbor,we can better retain the information of the dominant solution and improve the search efficiency.To verify the performance of the proposed generation strategy in solving the classification problem,twelve datasets are used in experiment.Experimental results show that the new generation strategy can improve the performance of BSO algorithm in solving classification problems.

Hybrid Support Vector Machines-Based Multi-fault Classification

Support Vector Machines （SVM） is a new general machine-learning tool based on structural risk minimization principle. This characteristic is very signific ant for the fault diagnostics when the number of fault samples is limited. Considering that SVM theory is originally designed for a two-class classification, a hybrid SVM scheme is proposed for multi-fault classification of rotating machinery in our paper. Two SVM strategies, 1-v-1 （one versus one） and 1-v-r （one versus rest）, are respectively adopted at different classification levels. At the parallel classification level, using l-v-1 strategy, the fault features extracted by various signal analysis methods are transferred into the multiple parallel SVM and the local classification results are obtained. At the serial classification level, these local results values are fused by one serial SVM based on 1-v-r strategy. The hybrid SVM scheme introduced in our paper not only generalizes the performance of signal binary SVMs but improves the precision and reliability of the fault classification results. The actually testing results show the availability suitability of this new method.

Balance between education and game-play— A case study of design and optimization of the mini-game book classification

Purpose： Based on our experience of designing and testing a computer-based game for teaching undergraduate students information literacy （IL） concepts and skills, this paper summarizes the basic strategies for striking a balance between education and entertainment for the designers of quality IL games. Design/methodology/approach： The project team recruited 10 college students to play the game and post-game group interviews revealed problems and optimization priorities. The optimized game was tested among 50 college students. Based on a comparison of testing results of the two versions of the game, basic strategies for designing quality 1L games were summarized. Findings： The following 5 basic strategies can effectively promote combination of education and entertainment： l） using adventure games to enhance gaming experience, 2） plotting an intriguing story to attract players, 3） motivating players to engage in game play with game components such as challenge, curiosity, fantasy and control, 4） presenting learning materials through game props, and 5） assigning players tasks to be completed with subject knowledge. Research limitations： The 5 basic strategies have been tested only in the development process of one game, and the book classification knowledge in the mini-game is limited to the 22 major categories of the Chinese Library Classification. Practical implications： University libraries may refer to our experience to design and utilize educational games to promote the IL education for college students. Originality value： Few empirical studies tested and summarized strategies for combining learning and fun in the design of IL games for university students. The 5 strategies, which are summarized in the process of design and optimization of the mini-game book classification, are valuable for other designers of IL games.

Global Strategies to Combat Antimicrobial Resistance: A One Health Perspective

Background: Antimicrobial resistance (AMR) is a global health challenge that has escalated due to the inappropriate use of antimicrobials in humans, animals, and the environment. Developing and implementing strategies to reduce and combat AMR is critical. Purpose: This study aimed to highlight some global strategies that can be implemented to address AMR using a One Health approach. Methods: This study employed a narrative review design that included studies published from January 2002 to July 2023. The study searched for literature on AMR and antimicrobial stewardship (AMS) in PubMed and Google Scholar using the 2020 PRISMA guidelines. Results: This study reveals that AMR remains a significant global public health problem. Its severity has been markedly exacerbated by inappropriate use of antimicrobials in humans, animals, and the broader ecological environment. Several strategies have been developed to address AMR, including the Global Action Plan (GAP), National Action Plans (NAPs), AMS programs, and implementation of the AWaRe classification of antimicrobials. These strategies also involve strengthening surveillance of antimicrobial consumption and resistance, encouraging the development of new antimicrobials, and enhancing regulations around antimicrobial prescribing, dispensing, and usage. Additional measures include promoting global partnerships, combating substandard and falsified antimicrobials, advocating for vaccinations, sanitation, hygiene and biosecurity, as well as exploring alternatives to antimicrobials. However, the implementation of these strategies faces various challenges. These challenges include low awareness and knowledge of AMR, a shortage of human resources and capacity building for AMR and AMS, in adequate funding for AMR and AMS initiatives, limited laboratory capacities for surveillance, behavioural change issues, and ineffective leadership and multidisciplinary teams. Conclusion: In conclusion, this study established that AMR is prevalent among humans, animals, and the environment. Successfully addressing AMR calls for a collaborative, multifaceted One Health approach. Despite this, some gaps remain effectively implementing strategies currently recommended to combat AMR. As a result, it is essential to reinforce the strategies that are deployed to counter AMR across the human, animal, and environmental sectors.

多区域注意力的细粒度图像分类网络

目前细粒度图像分类的难点在于如何精准定位图像中高度可辨的局部区域以及其他辅助判别特征。提出一种多区域注意力的细粒度图像分类网络来解决这个问题。首先使用Inception-V3对图像特征进行提取,通过重复使用注意力擦除的方法使模型关注次要特征;然后通过背景去除以及上采样的方法获取图像更精准的局部图像,对提取到的局部特征进行位置统计,并以矩形框的方式获取图像整体,减少细节信息丢失;最后对局部与整体图像进行更加细致的学习。此外,设计联合损失函数,通过动态平衡难易样本和缩小类内差距的方法改善模型的识别效果。实验结果表明,该方法在公开的细粒度图像数据集CUB-200-2011、Stanford-Cars和FGVC-Aircraft上的准确率分别达到89.2%、94.8%、94.0%,相较于对比方法性能更优。

风险动物园(Ⅶ):灰犀牛风险分类体系、演化机制与管控策略

作为风险管理领域的新兴概念和隐喻理论的重要组成部分,风险动物园中灰犀牛的本质属性和风险特征被广泛研究和应用。为防范化解新时代的风险挑战,加强灰犀牛相关研究的系统梳理和内涵挖掘,通过融合现有理论模型和借鉴风险分类方法,开展灰犀牛风险的类型体系构建、演化机制探析与管控策略研究。结果表明:灰犀牛风险作为客观环境因素和主观认知影响交互作用下的产物,按照特征维、认知维和状态维以及时间视角、空间视角和主体视角,把灰犀牛风险划分为隐藏灰犀牛、静止灰犀牛和冲撞灰犀牛;根据风险感知、风险距离和风险框架影响因素的作用机制,评估每类灰犀牛风险的发展态势;进而构建复杂系统内“累积—波动—突变”的演化框架,提出风险管控策略,提高认知能力察觉不易感知的隐藏灰犀牛;强化创新思维化解常态共存的静止灰犀牛;打造敏捷模式应对突然爆发的冲撞灰犀牛。

基于承载力—内生力—潜力的湖北省乡镇分类及规划策略研究

因地施策是乡镇级国土空间规划的要求,也是乡镇高质量发展的必然选择。以湖北省为例,以乡镇为基本单元,基于承载力—内生力—潜力分析乡镇发展情况,从区位条件、自然资源、社会经济、设施配套、地方特色5个维度34项指标构建乡镇分类指标体系,利用主客观叠加组合分析法赋值,通过自然间断点分级法进行乡镇发展类型划分。根据乡镇的综合优势和特色优势,将其分为“五型四类”,即城郊服务类、重点发展类、优化提升类、保留改善类以及生态资源型、旅游带动型、现代农业型、工业主导型、均衡发展型,并对各类乡镇的发展特点和空间格局进行对比分析。最后,针对综合分类和特色分类形成组合分类,提出相应的规划策略,为乡镇的分类评价和发展规划提供科学依据。

基于多节点策略的雷达智能干扰调制类型识别技术研究

采用认知雷达架构有助于实现雷达抗干扰技术的智能化程度提升。针对认知抗干扰技术领域中雷达对电磁干扰环境的感知问题,文中提出了一种基于深度学习的多节点干扰调制类型识别方法。该方法针对雷达信号处理的不同节点,如数字波束形成、自适应副瓣对消、脉冲压缩前后以及动目标检测之后,采用多个节点的时频平面和距离多普勒平面作为干扰信号的联合特征提取对象,建立了基于深度学习的多节点干扰识别策略模型,以提高多种干扰场景下的干扰识别正确率。为了提升干扰特征的提取能力和网络的训练效率,将用于干扰识别的深度学习算法在卷积神经网络的基础上引入了注意力机制和残差网络,建立了针对多节点策略下的干扰类型识别网络结构,实现了对多种不同干扰场景下的干扰类型识别。仿真结果表明:在单一干扰场景下,当干噪比为14 dB时,所提算法的干扰识别准确率可达92%;在多干扰场景下,所提算法在不同节点策略的加持下,识别准确率可达90%。

数字技术赋能生态文明建设:以城市生活垃圾分类为例

营建低废洁净的宜居环境,仰赖于居民生活方式行为习惯的转变。这一转变的促成既需要建立生态文明制度给予居民以外在约束,也需要强化生态文明道德来提升居民的内在动力。借助于演化博弈分析法来研究垃圾分类与投放行为在社区邻里间的相互影响与演化,并以仿真来呈现垃圾投放行为的深化路径、绿色行为能否成为普遍的行为选择。数字技术在城市生活垃圾管理上的深化应用,能降低监管成本提升监管效率,也能加强邻里的连接强度,提升社区融合程度,进而强化居民生态道德。研究结果显示,传统社区模式下封闭型社区的垃圾分类制度实施最终演化成理想的稳态结果,而在开放型社区绿色行为的转型难以实现。数字社区模式下封闭型社区和开放型社区,民众分类投放的“绿色低废”的行为将演变为普遍行为,数据技术能助力于绿色生活方式成风化俗。

基于混合策略改进ASO-LSSVM的风险驾驶行为分类识别

为解决现有智能算法在优化支持向量机识别风险驾驶行为过程中收敛速率缓慢和误差较大的问题。首先,采用Tent映射取代原子搜索优化算法(ASO)种群初始化随机设置的方式,增加原子种群多样性和质量;其次,使用逐维小孔成像反向学习与柯西变异混合机制,提高原子个体择优位置的多样性,克服ASO算法易陷入局部最优和过早收敛的问题;最后,通过引入自适应变螺旋搜寻策略改进原子个体位置更新过程,以提升ASO算法的全局搜索能力,实现全局搜索和局部开发间关系的有效平衡,缓解ASO算法易陷入局部最优和收敛精度不足的问题。以上海北横通道出口匝道车辆轨迹数据为输入,使用混合策略改进ASO算法寻优求解最小二乘支持向量机(LSSVM)参数,构建基于混合策略改进原子搜索优化最小二乘支持向量机IASO-LSSVM的快速路出口匝道风险驾驶行为分类识别模型。数值仿真实验结果表明:IASO算法在12个基准测试函数数值仿真结果的平均值、标准差、最佳适应度和最差适应度等方面均更接近最佳优化值。IASO-LSSVM模型相较于ASO-LSSVM和LSSVM等模型的风险驾驶行为分类识别结果误差指标正确率、精确率、召回率和F1值分别增加11.5~24.5、14.1~29.0、15.1~28.6和14.7~31.2个百分点,且在不同类型风险驾驶行为识别结果中误差变化范围最小。IASO算法参数寻优求解精度和收敛速率优于ASO算法,且IASO-LSSVM模型可用于不同类型风险驾驶行为精准识别,可为车辆行驶轨迹状态合理判别,制定风险驾驶行为预警防控措施提供数据支撑与理论依据。

基于改进SSA-DBN的质子交换膜燃料电池水故障智能分类方法

为了实现质子交换膜燃料电池(PEMFC)系统水故障的高效快速分类,提出了基于改进麻雀搜索算法(SSA)优化深度置信网络(DBN)的PEMFC故障分类方法。采用归一化处理消除故障数据参数之间量纲不同的影响,使用核主成分分析对数据进行故障特征提取,有效地缩减了原始数据维度,降低了运算复杂度,并避免低贡献度数据对故障分类造成干扰。引入柯西-高斯变异策略改进SSA,并利用SSA对DBN进行参数寻优,确定网络结构,通过优化后的DBN实现对PEMFC水故障的快速分类。对3 000组PEMFC水故障数据进行测试,结果表明:所提方法可以快速准确地识别PEMFC的正常状态、膜干故障、水淹故障3种健康状态;总体的分类准确率为98.67%,运算时间为0.89 s,相比支持向量机、概率神经网络方法,所提方法的故障分类精度分别提升了4%、3.34%,运算时间分别减少了15.35、0.35 s。

基于特征交互的层次分类在线流特征选择

在开放动态环境下的分类学习任务中,数据特征空间具有动态性,标记空间存在层次化结构.现有的层次分类在线流特征选择算法可以选择较优的特征子集,但这些算法忽略了特征之间存在的交互作用.基于此,提出了一种基于特征交互的层次分类在线流特征选择算法.首先,设计了一种基于层次邻域依赖度去判断特征交互的计算方法;其次,针对层次化结构数据,根据层次结构中不同节点间的兄弟关系定义邻域粗糙集模型;最后,设计了具有在线重要性分析、在线冗余性分析以及在线交互性分析的层次分类在线流框架,用于选择强相关和存在交互作用的特征子集.在6个层次数据集上的实验验证了所提算法具有较优的综合性能.

基于三生空间的川西林盘空间评估及建设策略研究--以都江堰市天马镇为例

川西林盘独特的宅、林、水、田景观格局是成都平原农耕文明的结晶。当前研究大多根据已有认知经验开展分析,较少结合空间分析评估对林盘保护提出建设策略。因此,本文以都江堰市天马镇林盘为例,首先,运用景观指数和空间分析方法,刻画研究区林盘的空间格局特征;其次,采用构建的三生空间评价指标体系对林盘生产、生活、生态空间的综合度和协调度进行评估,并识别出不同的林盘类型;最后,结合各类林盘类型的自然、经济和社会特征提出相应的未来发展建设策略。结果表明:天马镇林盘具有显著的规模不均衡性、空间异质性及景观破碎性特征;针对评价所得四种林盘类型,对天马镇林盘提出重点保护、原址修复和迁移合并三种空间建设策略。研究成果可对当地政府制定科学的林盘空间优化对策、保护措施提供科学参考。

求解约束优化问题的改进蛇优化算法

结合外点罚函数法与改进蛇优化算法求解约束优化问题,得到一种新的求解约束优化问题的算法WDFSO。算法WDFSO首先通过外点罚函数法将约束优化问题转化为一系列界约束优化问题,然后运用基于变异质心的对立学习策略与种群分类策略改进的蛇优化算法对所得界约束优化问题进行求解,进而获得所求约束优化问题的解。为验证算法WDFSO的有效性,选取CEC2006中19个标准约束优化问题进行数值实验,并使用Wilcoxon秩和检验来证明算法的显著性。实验结果表明,与对比算法相比,算法WDFSO求解约束优化问题具有更高的收敛精度和更好的稳定性。最后应用算法WDFSO求解两个工程约束优化问题,结果表明算法WDFSO求解性能更好。

A Synthesis of the Studies on Vocabulary Learning Strategies in ESL and EFL Learning

Vocabulary is the core of language learning.The efficiency of vocabulary learning may influence language learning outcome.Vocabulary learning strategies can help the learners to learn vocabulary more efficiently,and they are very useful and beneficial for language learners.This paper synthesizes the research on vocabulary learning strategies in ESL and EFL learning,including the studies on the definitions of vocabulary learning strategies,classifications of vocabulary learning strategies and previous main research areas.The previous research areas mainly involve in the following aspects:patterns and taxonomies of vocabulary learning strategies,gender difference and vocabulary learning strategies,language proficiency and vocabulary learning strategies,major or disciplinary difference and vocabulary learning strategies,vocabulary learning strategies used frequently and effectively,vocabulary learning strategies and learning outcomes,and vocabulary learning strategy training.

基于多目标优化的燃料电池汽车实时能量管理策略

为了降低混合动力系统的燃料消耗并延缓动力元件的老化,提出了一种基于多目标优化和路况分类的能量管理策略(EMS)。首先,构建了燃料电池与锂电池的电气模型,并引入了等效氢耗模型和燃料电池老化模型。其次,设计了基于规则的多模式EMS,在此基础上,为了进一步降低系统的等效氢耗,并延长其使用寿命,基于多目标白鲸算法(MOBWO)对EMS参数进行优化。再次,为了使所设计的EMS适用于不同的路况,提出了基于长短期记忆网络(LSTM)的驾驶路况实时分类方法,旨在根据分类结果切换EMS的控制参数以达到最优效果。最后,在仿真平台上对所提算法进行分析。结果表明:与基于规则的方法相比,所提方法氢耗量降低了2.3%,燃料电池的老化程度降低了1.02%,验证了所提EMS能够有效降低混合系统的燃料消耗,并且能够延缓燃料电池老化,从而提升了系统的经济性和耐久性。

数据与SEIR模型双驱动下COVID-19疫苗的空间分配策略

为提出一个适合疫情中后期新冠疫苗空间分配的新框架,通过收集2022年11月13日至28日国内31个省会城市的新冠疫情数据,求解传染病模型的逆问题得到初始潜伏人数,借助K-means算法对城市风险等级进行分类,进而根据风险分级后的城市建立疫苗分配模型,并通过遗传算法求解得到最优分配策略。数据结果表明,31个省会城市可分为5个疫情风险等级区域,风险等级越高的地区,疫苗在待接种人群的覆盖比例越高;在降低累计感染人数方面,所提出的分配策略优于地区人数比例分配策略和风险等级优先全覆盖分配策略。

家具图像阅读与设计项目研讨的进阶式教学

家具图像阅读能力是一种体现读者认知维度和水平的综合能力,是提升学生家具设计水平的基础。为了培养家具设计专业学生以图像为认知支架,提出问题、分析问题、解决问题的能力,以“学校宿舍用双层床安全设计”为题,借助布鲁姆认知目标分类系统和威廉姆斯教学策略,实施家具图像阅读与设计项目研讨进阶式教学,并形成一种创造性教学模式。结果表明,研讨进阶式教学项目有助于调动学生学习主动性,促进学生知识建构,培养学生批判性思维,提升学生创新能力。这种创造性教学模式对家具设计专业课程教学和学生综合能力培养是一种有益的探索。

主体功能区战略视角下广东省县域城镇化发展路径研究

推进以县城为重要载体的城镇化是新时期推进县域城镇化的重要切入点和关键抓手,对助推实现城乡融合和区域协调发展具有重大战略意义。本研究选取广东省57个县(市)作为研究对象,从落实主体功能区战略和完善主体功能区制度的视角,揭示了广东省县域主体功能区的重要特征,以及广东省县域城镇化、县域产业发展的情况和问题,并以底线优先、发展赋能、政策叠加的原则和思路,对衔接主体功能区战略的广东省县城进行了归类,对不同类别县城以产业为牵引推进城镇化的路径提出了具体对策建议,以期为支撑广东深入推进“百县千镇万村高质量发展工程”、为其他省市推进县域城镇化发展提供有益参考。