Multi-Valued Neuron with Sigmoid Activation Function for Pattern Classification

Multi-Valued Neuron (MVN) was proposed for pattern classification. It operates with complex-valued inputs, outputs, and weights, and its learning algorithm is based on error-correcting rule. The activation function of MVN is not differentiable. Therefore, we can not apply backpropagation when constructing multilayer structures. In this paper, we propose a new neuron model, MVN-sig, to simulate the mechanism of MVN with differentiable activation function. We expect MVN-sig to achieve higher performance than MVN. We run several classification benchmark datasets to compare the performance of MVN-sig with that of MVN. The experimental results show a good potential to develop a multilayer networks based on MVN-sig.

膳食模式与老年人群认知功能关系的研究进展

目的:比较国内外对认知功能有益或有害的膳食模式特点,为引导科学饮食、开发适用于我国的膳食模式提供参考,以便在我国中老年人群中推广。方法:对国内外探讨使用后验法和降秩回归法得到的膳食模式与认知功能或痴呆风险关系的研究进行综述。结果:研究发现,以植物性食物为特征的膳食模式与更好的认知表现相关,欧美国家的地域特色膳食模式与较差的认知功能相关。结论:有利于老年人群的认知功能的膳食模式需要包含丰富的食物种类,以蔬菜、水果、谷薯类、杂豆类、蛋类、乳品、畜禽肉和水产品作为主要的食物构成,避免红肉、脂肪、精制谷物和糖的大量摄入,并选择健康的烹饪方式。

基于加权Ripley's K-function的多尺度景观格局分析——以江苏盐城滨海湿地为例

以1987,1992,1997,2002,2007年的遥感影像为例,首次尝试使用加权Ripley's K-function的多尺度格局分析方法,计算了20年来景观异质性在江苏盐城滨海湿地的时间变化和空间分布趋势。通过对研究区的样带划分以及景观类型的点状化处理,建立滨海湿地样带图层和1987—2007年间各类型景观的点格局数据库,从而分析滨海湿地不同类型景观的空间聚集特征变化。基于加权Ripley's K-function的计算表明,在各级空间尺度和时间变化上,各类型湿地的斑块都呈现出空间聚集分布状态,且1987年以来,不同湿地类型的聚集空间特征尺度和空间分布强度均出现了大幅的增减变化,除互花米草滩之外的自然湿地的聚集空间特征尺度和强度都有明显下降甚至少到无法被检测到,而人工湿地却呈现聚集特征尺度和强度的双增长,且该聚集程度有逐渐增强的趋势。分析表明,既考虑样点的空间位置信息又考虑样点分布范围的加权Ripley's K-function方法能很好地表征湿地景观在多尺度上的变异,且与传统空间景观指数等分析方法的结论在一定程度上保持一致。

人口老龄化视阈下职业教育的价值重构、功能定位与实践样态

随着人口老龄化问题日益凸显,探讨如何有效化解老龄化带来的社会问题,成了时代发展的必答题。通过调查S市H区老年人生活与再就业状况,综合运用logistic回归模型等方法,深度剖析影响老年人再就业意愿的多重因素,发现有技能特长、年龄较小、学历较高与老年人再就业率成正相关。沿着老年人职业教育的价值重构、功能定位与实践样态进行三重进路分析,并围绕提高个体生活质量、适应外界社会环境变化、追寻积极个体存在意义、做品德高尚和有用有为的老年人等多维度提出了老年人职业教育的价值重构,以期为我国积极应对人口老龄化对策提供有益参考。

基于改进YOLOv8s的花色布疵点检测算法

针对花色布背景图案复杂、部分疵点目标小、与背景分离难度大,及其带来的自动化实时检测挑战等问题,提出一种基于改进YOLOv8s的花色布疵点检测算法。为增强网络检测疵点小目标的能力,引入BiFPN(Bi⁃directional Feature Pyramid)特征融合网络,充分融合语义信息和位置信息,增强算法区分疵点和背景的能力;同时,注意到疵点小目标中低质量样本对检测结果的影响,引入WIoU v3损失函数,抑制训练过程中低质量样本产生的有害梯度;最后,引入FasterBlock模型改进原始模型neck中的C2f模块,降低模型整体的参数量。结果表明:改进后的YOLOv8s模型在整个数据集的mAP@0.5∶0.95上可达59.6%,比原YOLOv8s模型提高了2.8个百分点;小目标的APs@0.5∶0.95可达45.1%,比原YOLOv8s模型提高了8.3个百分点;改进后模型参数量为10.557 M,检测速度可达131.6帧/s。认为:改进的YOLOv8s有效提升了沾污、花毛等小目标的检测效果。

Alteration of functional connectivity in patients with Alzheimer’s disease revealed by resting-state functional magnetic resonance imaging

The main symptom of patients with Alzheimer’s disease is cognitive dysfunction. Alzheimer’s disease is mainly diagnosed based on changes in brain structure. Functional connectivity reflects the synchrony of functional activities between non-adjacent brain regions, and changes in functional connectivity appear earlier than those in brain structure. In this study, we detected resting-state functional connectivity changes in patients with Alzheimer’s disease to provide reference evidence for disease prediction. Functional magnetic resonance imaging data from patients with Alzheimer’s disease were used to show whether particular white and gray matter areas had certain functional connectivity patterns and if these patterns changed with disease severity. In nine white and corresponding gray matter regions, correlations of normal cognition, early mild cognitive impairment, and late mild cognitive impairment with blood oxygen level-dependent signal time series were detected. Average correlation coefficient analysis indicated functional connectivity patterns between white and gray matter in the resting state of patients with Alzheimer’s disease. Functional connectivity pattern variation correlated with disease severity, with some regions having relatively strong or weak correlations. We found that the correlation coefficients of five regions were 0.3–0.5 in patients with normal cognition and 0–0.2 in those developing Alzheimer’s disease. Moreover, in the other four regions, the range increased to 0.45–0.7 with increasing cognitive impairment. In some white and gray matter areas, there were specific connectivity patterns. Changes in regional white and gray matter connectivity patterns may be used to predict Alzheimer’s disease;however, detailed information on specific connectivity patterns is needed. All study data were obtained from the Alzheimer’s Disease Neuroimaging Initiative Library of the Image and Data Archive Database.

Spatio-temporal point pattern analysis on Wenchuan strong earthquake

For exploring the aftershock occurrence process of the 2008 Wenchuan strong earthquake, the spatio-temporal point pattern analysis method is employed to study the sequences of aflershocks with magnitude M≥4.0, M≥4.5, and M≥5.0. It is found that these data exhibit the spatio-temporal clustering on a certain distance scale and on a certain time scale. In particular, the space-time interaction obviously strengthens when the distance is less than 60 km and the time is less than 260 h for the first two aftershock sequences; however, it becomes strong when the distance scale is less than 80 km and the time scale is less than 150 h for the last aftershock sequence. The completely spatial randomness analysis on the data regardless of time component shows that the spatial clustering of the aftershocks gradually strengthens on the condition that the distance is less than 60 km. The results are valuable for exploring the occurrence rules of the Wenchuan strong earthquake and for predicting the aftershocks.

Statistical Classification Using the Maximum Function

The maximum of k numerical functions defined on , , by , ??is used here in Statistical classification. Previously, it has been used in Statistical Discrimination [1] and in Clustering [2]. We present first some theoretical results on this function, and then its application in classification using a computer program we have developed. This approach leads to clear decisions, even in cases where the extension to several classes of Fisher’s linear discriminant function fails to be effective.

Changes in Transformation of Soil Organic C and Functional Diversity of Soil Microbial Community Under Different Land Uses

Changes in soil biological and biochemical properties under different land uses in the subtropical region of China were investigated in order to develop rational cultivation and fertilization management. A small watershed of subtropical region of China was selected for this study. Land uses covered paddy fields, vegetable farming, fruit trees, upland crops, bamboo stands, and forestry. Soil biological and biochemical properties included soil organic C and nutrient contents, mineralization of soil organic C, and soil microbial biomass and community functional diversity. Soil organic C and total N contents, microbial biomass C and N, and respiration intensity under different land uses were changed in the following order： paddy fields （and vegetable farming） 〉 bamboo stands 〉 fruit trccs （and upland）. The top surface （0-15 cm） paddy fields （and vegetable farming） were 76.4 and 80.8% higher in soil organic C and total N contents than fruit trees （and upland） soils, respectively. Subsurface paddy soils （15-30 cm） were 59.8 and 67.3% higher in organic C and total N than upland soils, respectively. Soil microbial C, N and respiration intensity in paddy soils （0-15 cm） were 6.36, 3.63 and 3.20 times those in fruit tree （and upland） soils respectively. Soil microbial metabolic quotient was in the order： fruit trees （and upland） 〉 forestry 〉 paddy fields. Metabolic quotient in paddy soils was only 47.7% of that in fruit tree （and upland） soils. Rates of soil organic C mineralization during incubation changed in the order： paddy fields 〉 bamboo stands 〉 fruit trees （and upland） and soil bacteria population： paddy fields 〉 fruit trees （and upland） 〉 forestry. No significant difference was found for fungi and actinomycetes populations. BIOLOG analysis indicated a changing order of paddy fields 〉 fruit trees （and upland） 〉 forestry in values of the average well cell development （AWCD） and functional diversity indexes of microbial community. Results also showed that the conversion from paddy fields to vegetable farming for 5 years resulted in a dramatic increase in soil available phosphorus content while insignificant changes in soil organic C and total N content due to a large inputs of phosphate fertilizers. This conversion caused 53, 41.5, and 41.3% decreases in soil microbial biomass C, N, and respiration intensity, respectively, while 23.6% increase in metabolic quotient and a decrease in soil organic C mineralization rate. Moreover, soil bacteria and actinomycetes populations were increased slightly, while fungi population increased dramatically. Functional diversity indexes of soil microbial community decreased significantly. It was concluded that land uses in the subtropical region of China strongly affected soil biological and biochemical properties. Soil organic C and nutrient contents, mineralization of organic C and functional diversity of microbial community in paddy fields were higher than those in upland and forestry. Overuse of chemical fertilizers in paddy fields with high fertility might degrade soil biological properties and biochemical function, resulting in deterioration of soil biological quality.

Application study of image segmentation methods on pattern recognition in the course of wood across-compression

Image segmentation is one of important steps on pattern recognition study in the course of wood across-compression. By comparing and studying processing methods in finding cell space and cell wall, this paper puts forward some image segmentation methods that are suitable for study of cell images of wood crossgrained compression. The method of spline function fitting was used for linking edges of cell, which perfects the study of pattern recognition in the course of wood across-compression.

Effect of the Illuminating Point Spread Function and Roughness on the Speckle Correlation Field

Under conditions differing from those subjected for central limit theorem, the spatial autocorrelation function of speckle pattern resulting from illuminated rough surface is investigated. Its dependence on different illuminating apertures and the average of the roughness heights is presented theoretically and experimentally. The experiments were carried out using a set of circular and square apertures having different sizes. The results indicate that, increasing the size of the illuminating aperture leads to a decrease in the width of the main lobe of the spatial autocorrelation function.

PATTERN MINING AND DISCOVERY ORIENTED TO ARTIFICIAL LIFE

１ＩＮＴＲＯＤＵＣＴＩＯＮＰａｔｅｒｎｍｉｎｉｎｇａｎｄｄｉｓｃｏｖｅｒｙｉｓｏｎｅｏｆｔｈｅｅｍｅｒｇｉｎｇｂｒａｎｃｈｅｓｏｆＫＤＤ（ＫｎｏｗｌｅｄｇｅＤｉｓｃｏｖｅｒｙｉｎＤａｔａｂａｓｅ）ｗｈｉｃｈｉｓａｐｒｏｍｉｓｉｎｇｆｉｅｌｄｏｆｔｈ...

Functional Division Analysis and Optimization Design of Nanling Nature Reserve

Nanling National Nature Reserve is established on the basis of 5 provincial nature reserves, and the original functional zoning is maintained. This paper explored problems of the reserve in division range and functional zoning, and gavea pertinent adjustment scheme. The adjustment results are(1) cutting out the scattered areas in neighborhood, and fi lling in the internal split parts;(2) establishing corridors between the reserve and neighboring nature reserves;(3) redividing the functional zones.

Spatio-Temporal Patterns of Wildlife Road Mortality in Golestan National Park-North East of Iran

Nowadays, wildlife road mortality is acknowledged as a main source of threatening long-term survival of wildlife. This paper as the first to analysis wild life vehicle collisions in Iran, aims to reconstruct and interpret the spatio-temporal patterns of WVCs on Asiaei highway in Golestan National Park (GNP). With the collaboration of environmental protection department of GNP, we identified about 1900 WVC Records involving 34 different species of mammals, birds, reptiles and amphibians between 2004 and 2013. Mammals were involved in more than 50% of overall WVCs, among which wild boar (Sus scrofa), Golden Jackal (Canis aureus), Red Fox (Vulpes vulpes), hedgehog (Erinaceus concolor), stone marten (Martes foina) and porcupine (Hystrix indica) were involved in more than 90% of mammals’ mortalities;So, we focused on analyzing spatio-temporal pattern of vehicle collisions of these six mammal species. During the study period, these species have undergone 95% increase in road mortalities, averagely. Detailed temporal analyses exhibited an increasing trend of road mortalities from spring to summer and then a reducing one to late winter. It was shown that a large number of collisions occurred in holiday periods when recreational trips considerably increased the traffic volume of Asiaei highway. Preliminary inspection of spatial patterns using Kernel density analysis revealed six collision hotspots, mostly located in the road bends with densely forested land cover on both sides;the promenades along the road seemed to play a significant role too. Scale dependency analyses of collision patterns, demonstrated clustering pattern at micro scales less than 10 km, randomness at meso scales 10 - 20 km and both regularity and clustering at macro scales more than 20 km. This paper suggests that road mortality of common species in GNP is a momentous issue, which needs to be considered by relevant governmental and public organizations. We also emphasize that the analyses of spatial and temporal patterns of WVCs are fundamentals to plan for mitigate wildlife road mortality.

An Autonomous Incremental Learning Algorithm for Radial Basis Function Networks

In this paper, an incremental learning model called Resource Allocating Network with Long-Term Memory (RAN-LTM) is extended such that the learning is conducted with some autonomy for the following functions: 1) data collection for initial learning, 2) data normalization, 3) addition of radial basis functions (RBFs), and 4) determination of RBF cen-ters and widths. The proposed learning algorithm called Autonomous Learning algorithm for Resource Allocating Network (AL-RAN) is divided into the two learning phases: initial learning phase and incremental learning phase. And the former is further divided into the autonomous data collection and the initial network learning. In the initial learning phase, training data are first collected until the class separability is converged or has a significant dif-ference between normalized and unnormalized data. Then, an initial structure of AL-RAN is autonomously determined by selecting a moderate number of RBF centers from the collected data and by defining as large RBF widths as possible within a proper range. After the initial learning, the incremental learning of AL-RAN is conducted in a sequential way whenever a new training data is given. In the experiments, we evaluate AL-RAN using five benchmark data sets. From the experimental results, we confirm that the above autonomous functions work well and the efficiency in terms of network structure and learning time is improved without sacrificing the recognition accuracy as compared with the previous version of AL-RAN.

Design Pattern Representation for Safety-Critical Embedded Systems

Design Patterns, which give abstract solutions to commonly recurring design problems, have been widely used in the software and hardware domain. As non-functional requirements are an important aspect in the design of safety-critical embedded systems, this work focuses on the integration of non-functional implications in an existing design pattern concept. We propose a pattern representation for safety-critical embedded application design methods by including fields for the implications and side effects of the represented design pattern on the non-functional requirements of the overall systems. The considered requirements include safety, reliability, modifiability, cost, and execution time.

PATTERN MAPPING ON BETA-SPLINE SURFACES——BILINEAR MAPPING METHOD

Texture pattern mapping is one of the most important techniques for high quality image syn- thesis. It can largely enhance the visual richness of raster-scan images. In this paper is presented a new method of mapping planar texture pattern onto beta-spline curved surfaces——bilinear mapping method which can map planar texture pattern onto curved surfaces with less distortion, and also can fulfill the geometric transformation of the texture pattern on the curved surfaces by operating the pattern win- dow. It is valuable to both CAD/CAM in artistic field and computer graphics.

基于多模型和权衡矩阵法的乡村生态-游憩景观安全格局重构

【目的】乡村旅游地的突出矛盾是人类旅游活动对生态环境的干扰,景观规划时需要权衡游憩功能与生态功能。探究乡村生态-游憩景观安全格局(landscape security pattern, LSP)重构路径,旨在促进乡村旅游与生态环境和谐、可持续发展,为乡村旅游地景观规划提供参考思路。【方法】以江苏省无锡市马山为例,综合运用InVEST、ArcSWAT、最小累积阻力(minimum cumulative resistance, MCR)多模型,分别模拟生态、游憩空间格局以及生态、游憩廊道,并划分景观单元;基于“权衡矩阵法”,比较各单元生态、游憩功能重要性,实现景观功能分区,识别权衡问题突出的区域及战略点,进而为各类景观功能区提出针对性的规划、建设及管控策略。【结果】将马山景观功能区划分为生态核心区、生态重要区、生态-游憩重点权衡区、生态-游憩次级权衡区、游憩核心区、游憩重要区、游憩协调区及弹性发展区8类,构建生态-游憩景观网络并识别出32处战略点。【结论】重构乡村生态-游憩LSP有利于区分不同类型景观功能区的建设侧重点与优先级,科学、精细化地管理、利用乡村景观资源,促进多功能景观的协调与增效。

小波分析概述:从信号处理的视角

本文从信号处理的视角对20世纪80年代末兴起的小波分析做一个通俗的概述.内容包括小波分析产生的基本问题背景和关键历史事件,小波分析的基本理论,及其对信号处理的意义和作用.希望透过本文能科普性地展示小波的基本理论以及为什么要建立这样的理论.除此以外,笔者还对小波分析之后时频分析领域所出现的经验模型分解方法和图信号处理的基本问题给出简要的介绍.作为中山大学百年校庆的约稿综述,本文对中山大学小波分析研究团队的基本阵容以及历年来所开展的重要学术事件进行了简单的回顾.

资源枯竭型城市生态质量时空格局与驱动力研究--以阜新市中心城区为例

为评价资源枯竭型城市转型期的生态质量时空演变特征,本文以阜新市中心城区为研究区域,基于Landsat遥感影像构建遥感生态指数(RSEI),以OSM数据和POI数据为基础识别城市功能区,借助参数最优地理探测器模型探究城市生态质量与城市功能区之间的复杂性关系。结果表明:①2004—2022年阜新市中心城区生态质量整体呈略微下降趋势、波动相对稳定,形成四周好、中间差的盆地状格局;②Hurst指数均值为0.67,强持续性最近邻指数值为146.53,生态质量未来变化趋势呈现均匀持续性,大部分区域变化趋势与过去相同,但局部地区有反持续性趋势;③识别出“31+1”类城市功能区,功能区生态质量由高到低依次为绿地、公共服务、工业、交通、居住、商业、矿区,且不同因子交互作用均会增强对生态质量的解释力。研究结果可为资源枯竭型城市转型期针对不同功能区的生态保护和恢复工作提供科学参考。