Deep Learning Social Network Access Control Model Based on User Preferences

A deep learning access controlmodel based on user preferences is proposed to address the issue of personal privacy leakage in social networks.Firstly,socialusers andsocialdata entities are extractedfromthe social networkandused to construct homogeneous and heterogeneous graphs.Secondly,a graph neural networkmodel is designed based on user daily social behavior and daily social data to simulate the dissemination and changes of user social preferences and user personal preferences in the social network.Then,high-order neighbor nodes,hidden neighbor nodes,displayed neighbor nodes,and social data nodes are used to update user nodes to expand the depth and breadth of user preferences.Finally,a multi-layer attention network is used to classify user nodes in the homogeneous graph into two classes:allow access and deny access.The fine-grained access control problem in social networks is transformed into a node classification problem in a graph neural network.The model is validated using a dataset and compared with other methods without losing generality.The model improved accuracy by 2.18%compared to the baseline method GraphSAGE,and improved F1 score by 1.45%compared to the baseline method,verifying the effectiveness of the model.

Multimodal Social Media Fake News Detection Based on Similarity Inference and Adversarial Networks

As social networks become increasingly complex, contemporary fake news often includes textual descriptionsof events accompanied by corresponding images or videos. Fake news in multiple modalities is more likely tocreate a misleading perception among users. While early research primarily focused on text-based features forfake news detection mechanisms, there has been relatively limited exploration of learning shared representationsin multimodal (text and visual) contexts. To address these limitations, this paper introduces a multimodal modelfor detecting fake news, which relies on similarity reasoning and adversarial networks. The model employsBidirectional Encoder Representation from Transformers (BERT) and Text Convolutional Neural Network (Text-CNN) for extracting textual features while utilizing the pre-trained Visual Geometry Group 19-layer (VGG-19) toextract visual features. Subsequently, the model establishes similarity representations between the textual featuresextracted by Text-CNN and visual features through similarity learning and reasoning. Finally, these features arefused to enhance the accuracy of fake news detection, and adversarial networks have been employed to investigatethe relationship between fake news and events. This paper validates the proposed model using publicly availablemultimodal datasets from Weibo and Twitter. Experimental results demonstrate that our proposed approachachieves superior performance on Twitter, with an accuracy of 86%, surpassing traditional unimodalmodalmodelsand existing multimodal models. In contrast, the overall better performance of our model on the Weibo datasetsurpasses the benchmark models across multiple metrics. The application of similarity reasoning and adversarialnetworks in multimodal fake news detection significantly enhances detection effectiveness in this paper. However,current research is limited to the fusion of only text and image modalities. Future research directions should aimto further integrate features fromadditionalmodalities to comprehensively represent themultifaceted informationof fake news.

METAGENOMICS COMBINED WITH HIGHTHROUGHPUT SEQUENCING REVEALS THE METHANOGENIC POTENTIAL OF FRESH CORN STRAW UNDER THERMOPHILIC AND HIGH OLR

Dry corn straw(DCS)is usually used in anaerobic digestion(AD),but fresh corn straw(FCS)has been given less consideration.In this study,the thermophilic AD of single-substrate(FCS and DCS)and co-digestion(straw with cattle manure)were investigated.The results show that when FCS was used as the single-substrate for AD,the methane production was 144 mL·g^(−1)·VS^(−1),which was 7.5%and 19.6%higher than that of single DCS and FCS with cattle manure,respectively.In addition,the structure of FCS was loose and coarse,which was easier to be degraded than DCS.At the hydrolysis and acidification stages,Clostridium_sensu_stricto_1,Clostridium_sensu_stricto_7 and Sporosarcina promoted the decomposition of organic matter,leading to volatile fatty acids(VFAs)accumulation.Methanosarcina(54.4%)activated multifunctional methanogenic pathways to avoid the VFAs inhibition,which was important at the CH_(4) production stage.The main pathway was hydrogenotrophic methanogenesis,with genes encoding formylmethanofuran dehydrogenase(K00200-K00203)and tetrahydromethanopterin Smethyltransferase(K00577-K00584).Methanosarcina also activated acetotrophic and methylotrophic methanogenesis pathways,with genes encoding acetyl phosphate(K13788)and methyl-coenzyme M reductase(K04480,K14080 and K14081),respectively.In the co-digestion,the methanogenic potential of FCS was also confirmed.This provides a scientific basis for regulating AD of crop straw.

拓扑切换的集群系统分布式分组时变编队跟踪控制

分布式编队控制是集群智能控制重要的基础性研究课题之一.拒止环境和多任务需求对编队控制提出了新的挑战.本文研究了高阶线性集群系统在拓扑切换情况下的分布式分组时变编队跟踪控制问题.在本文所提出的集群系统模型架构中,智能体被分为3类:虚拟领导者、分组领导者和跟随者.虚拟领导者为整个集群系统宏观运动提供参考轨迹或者跟踪指令.分组领导者一方面跟踪虚拟领导者所提供的轨迹或指令;另一方面通过与其他分组领导者之间的相互协作实现各分组之间的协同配合.跟随者实现对各自分组领导者的时变编队跟踪.在拓扑切换和外部扰动同时存在的情况下,基于智能体之间的局部有限邻居节点之间的相对信息反馈和滑模控制理论构造了分布式分组时变编队跟踪控制协议,并给出了控制协议中未知参数的求解算法,进而利用李雅普诺夫(Lyapunov)理论证明了集群系统在拓扑切换和外部扰动同时存在情况下的闭环稳定性.最后,数值仿真结果验证了本文提出的控制方法能够实现集群系统分组时变编队跟踪控制,并可应用在以无人机集群为代表的实际物理模型中.

Machine learning helps improve diagnostic ability of subclinical keratoconus using Scheimpflug and OCT imaging modalities

Purpose:To develop an automated classification system using a machine learning classifier to distinguish clinically unaffected eyes in patients with keratoconus from a normal control population based on a combination of Scheimpflug camera images and ultra-high-resolution optical coherence tomography(UHR-OCT)imaging data.Methods:A total of 121 eyes from 121 participants were classified by 2 cornea experts into 3 groups:normal(50 eyes),with keratoconus(38 eyes)or with subclinical keratoconus(33 eyes).All eyes were imaged with a Scheimpflug camera and UHR-OCT.Corneal morphological features were extracted from the imaging data.A neural network was used to train a model based on these features to distinguish the eyes with subclinical keratoconus from normal eyes.Fisher’s score was used to rank the differentiable power of each feature.The receiver operating characteristic(ROC)curves were calculated to obtain the area under the ROC curves(AUCs).Results:The developed classification model used to combine all features from the Scheimpflug camera and UHR-OCT dramatically improved the differentiable power to discriminate between normal eyes and eyes with subclinical keratoconus(AUC=0.93).The variation in the thickness profile within each individual in the corneal epithelium extracted from UHR-OCT imaging ranked the highest in differentiating eyes with subclinical keratoconus from normal eyes.Conclusion:The automated classification system using machine learning based on the combination of Scheimpflug camera data and UHR-OCT imaging data showed excellent performance in discriminating eyes with subclinical keratoconus from normal eyes.The epithelial features extracted from the OCT images were the most valuable in the discrimination process.This classification system has the potential to improve the differentiable power of subclinical keratoconus and the efficiency of keratoconus screening.

Formation process and properties of hydrogen-producing granular sludge in UASB reactor

The granular sludge of microbial fermentation systems includes various biomass-degrading enzymes and different microflora,which have significant impacts on the degradation of biomass and the stability of the system.An up-flow anaerobic sludge blanket(UASB)reactor was used to grow hydrogen-producing granular sludge.The results showed that the formation of the granular sludge underwent four stages,i.e.,flocculation of the sludge,formation of the flocculent sludge,swelling of the flocculent sludge,and formation of the granular sludge.The formed granular sludge mostly had regular spherical and ellipsoidal shapes with a fractal dimension of 2.08±0.4;the settling velocities were 0.84 cm/s to 1.96 cm/s in water,the porosity was 0.67-0.95.The shear sensitivity(Kss)of the granular sludge was 0.1152.The granular sludge had a culture cycle of approximately 70 d and a hydrogen yield of 1.09 mol H2/mol glucose.

Modeling of gonioscopic anterior chamber angle grades based on anterior segment optical coherence tomography

Background:To quantitatively assess anterior chamber angle(ACA)structure by anterior segment optical coherence tomography(AS-ocT)and develop a model to evaluate angle width as defined by gonioscopy.Methods:The ACAs of each quadrant were evaluated by gonioscopy,classified by the Scheie grading system,and assigned into one of the three grades:small angle(SA),moderate angle(MA),and large angle(LA).The eyes were imaged by AS-OCT,and ACA structural parameters including angle opening distance at the scleral spur(AODSS)and at 750μm anterior to the scleral spur(AOD750),trabecular-iris space area at 750μm anterior to the scleral spur(TISA750),and a newly defined parameter"light intersection distance"(LID),were measured.The ACA structural data were used to construct an ordered logistic regression model for assignment of ACAs to one of the three angle grades.The validity of the model was then tested.Results:A total of 169 quadrants from 53 subjects were included in the analysis,of which 111 quadrants were included in the modeling data and 58 in the testing data.In pairwise comparisons of SA,MA,and LA by ANOVA,the measured parameters were as follows:AOD750(0.174±0.060 vs.0.249±0.068 vs.0.376±0.114 mm;P<0.001),TISA750(0.075±0.035 vs.0.117±0.036 vs.0.181±0.062 mm^(2);P<0.001),and LID(0.300±0.187 vs.0.085±0.170 vs.0.122±0.156 mm;P<0.001).The ACA grading model based on LID showed a relatively high correction rate of 72.4%,and the model efficiency,calculated using the receiver operating characteristic,showed an area under the curve of 0.740.Weighted kappa statistics showed a good agreement for multiple ACA grades(0.772).Conclusions:The AS-OCT-based multiple ACA grades model was demonstrated as a non-contact approach for ACA assessment with high speed and high spatial resolution,providing guidance for diagnosis of angle closure.

Modeling of gonioscopic anterior chamber angle grades based on anterior segment optical coherence tomography

Background:To quantitatively assess anterior chamber angle(ACA)structure by anterior segment optical coherence tomography(AS-OCT)and develop a model to evaluate angle width as defined by gonioscopy.Methods:The ACAs of each quadrant were evaluated by gonioscopy,classified by the Scheie grading system,and assigned into one of the three grades:small angle(SA),moderate angle(MA),and large angle(LA).The eyes were imaged by AS-OCT,and ACA structural parameters including angle opening distance at the scleral spur(AODSS)and at 750μm anterior to the scleral spur(AOD750),trabecular-iris space area at 750μm anterior to the scleral spur(TISA750),and a newly defined parameter“light intersection distance”(LID),were measured.The ACA structural data were used to construct an ordered logistic regression model for assignment of ACAs to one of the three angle grades.The validity of the model was then tested.Results:A total of 169 quadrants from 53 subjects were included in the analysis,of which 111 quadrants were included in the modeling data and 58 in the testing data.In pairwise comparisons of SA,MA,and LA by ANOVA,the measured parameters were as follows:AOD750(0.174±0.060 vs.0.249±0.068 vs.0.376±0.114 mm;P<0.001),TISA750(0.075±0.035 vs.0.117±0.036 vs.0.181±0.062 mm^(2);P<0.001),and LID(−0.300±0.187 vs.-0.085±0.170 vs.0.122±0.156 mm;P<0.001).The ACA grading model based on LID showed a relatively high correction rate of 72.4%,and the model efficiency,calculated using the receiver operating characteristic,showed an area under the curve of 0.740.Weighted kappa statistics showed a good agreement for multiple ACA grades(0.772).Conclusions:The AS-OCT-based multiple ACA grades model was demonstrated as a non-contact approach for ACA assessment with high speed and high spatial resolution,providing guidance for diagnosis of angle closure.