Evaluating the adversarial robustness of classification algorithms in machine learning is a crucial domain.However,current methods lack measurable and interpretable metrics.To address this issue,this paper introduces ...Evaluating the adversarial robustness of classification algorithms in machine learning is a crucial domain.However,current methods lack measurable and interpretable metrics.To address this issue,this paper introduces a visual evaluation index named confidence centroid skewing quadrilateral,which is based on a classification confidence-based confusion matrix,offering a quantitative and visual comparison of the adversarial robustness among different classification algorithms,and enhances intuitiveness and interpretability of attack impacts.We first conduct a validity test and sensitive analysis of the method.Then,prove its effectiveness through the experiments of five classification algorithms including artificial neural network(ANN),logistic regression(LR),support vector machine(SVM),convolutional neural network(CNN)and transformer against three adversarial attacks such as fast gradient sign method(FGSM),DeepFool,and projected gradient descent(PGD)attack.展开更多
Confusing object detection(COD),such as glass,mirrors,and camouflaged objects,represents a burgeoning visual detection task centered on pinpointing and distinguishing concealed targets within intricate backgrounds,lev...Confusing object detection(COD),such as glass,mirrors,and camouflaged objects,represents a burgeoning visual detection task centered on pinpointing and distinguishing concealed targets within intricate backgrounds,leveraging deep learning methodologies.Despite garnering increasing attention in computer vision,the focus of most existing works leans toward formulating task-specific solutions rather than delving into in-depth analyses of methodological structures.As of now,there is a notable absence of a comprehensive systematic review that focuses on recently proposed deep learning-based models for these specific tasks.To fill this gap,our study presents a pioneering review that covers both themodels and the publicly available benchmark datasets,while also identifying potential directions for future research in this field.The current dataset primarily focuses on single confusing object detection at the image level,with some studies extending to video-level data.We conduct an in-depth analysis of deep learning architectures,revealing that the current state-of-the-art(SOTA)COD methods demonstrate promising performance in single object detection.We also compile and provide detailed descriptions ofwidely used datasets relevant to these detection tasks.Our endeavor extends to discussing the limitations observed in current methodologies,alongside proposed solutions aimed at enhancing detection accuracy.Additionally,we deliberate on relevant applications and outline future research trajectories,aiming to catalyze advancements in the field of glass,mirror,and camouflaged object detection.展开更多
Recently,in Northeast Asian situation there have occurred escalation of tension and intensification of antagonism seldom seen before.At the same time,however,there are also explorations and signs of restarting dialogu...Recently,in Northeast Asian situation there have occurred escalation of tension and intensification of antagonism seldom seen before.At the same time,however,there are also explorations and signs of restarting dialogue and cooperation that have stagnated since long ago.Development trends of the situation are confusing to the eye,but reconciliation and cooperation remain the way-out for展开更多
Lactic acid bacteria(LAB)were probiotics that produced a variety of metabolites.As a probiotic,exopolysaccharide(EPS)from LAB was widely used in the fields of food and medicine.Morphological observation,API 50 CHL tes...Lactic acid bacteria(LAB)were probiotics that produced a variety of metabolites.As a probiotic,exopolysaccharide(EPS)from LAB was widely used in the fields of food and medicine.Morphological observation,API 50 CHL test and 16S rDNA were used to identify the species characteristics of EPS producing LAB,and to explore the strain tolerance characteristics.The results showed that the strain P2 was Weissella confusa and named W.confusa P2.The survival rate of this strain was more than 60%in the range of 0~0.25 g·L^(-1) bile salt concentration,and it could survive well at pH 4.0~7.5 and NaCl concentration 0~4 g·L^(-1).Moreover,the EPS content of W.confusa P2 reached a maximum of 16.06±0.33 g·L^(-1) when the fermentation time was 36 h.W.confusa P2 may be a potential candidate strain in the field of food fermentation and possibly probiotics.Therefore,as a strain with tolerance and high EPS yield,W.confusa P2 not only had significant application potential in the food industry,but also provided new resources and directions for the development of probiotic preparations.展开更多
The term“nipple confusion”accurately describes the confusion newborns experience between their mother’s nipple and an artificial nipple during feeding.Specifically,it refers to the feeding habits infants develop ba...The term“nipple confusion”accurately describes the confusion newborns experience between their mother’s nipple and an artificial nipple during feeding.Specifically,it refers to the feeding habits infants develop based on their initial feeding experiences after birth.Infants accustomed to the maternal nipple often resist bottle-feeding;conversely,those accustomed to bottle-feeding may reject the maternal nipple.This confusion is particularly common among infants receiving mixed feeding.展开更多
Nipple confusion,a term that accurately describes the confusion that occurs between the mother’s nipple and the artificial teat during feeding in newborns.Specifically,it refers to the fact that babies develop specif...Nipple confusion,a term that accurately describes the confusion that occurs between the mother’s nipple and the artificial teat during feeding in newborns.Specifically,it refers to the fact that babies develop specific breastfeeding habits after birth,based on their initial feeding experience.For babies who are accustomed to their mother’s nipple,they tend to show resistance to bottle feeding;on the contrary,those who have adapted to bottle feeding may refuse to accept their mother’s nipple.This confusion is particularly common among mixed-feeding babies.展开更多
The leaf epidermis of Japanese honeysuckle （Lonicera japonica Thunb.） and Wild Honeysuckle （Lonicera confusa） in the genus of Flos Lonicerae were mainly observed by scanning electron microscopes （SEM） to study t...The leaf epidermis of Japanese honeysuckle （Lonicera japonica Thunb.） and Wild Honeysuckle （Lonicera confusa） in the genus of Flos Lonicerae were mainly observed by scanning electron microscopes （SEM） to study the characteristics of stomata, trichomes and dermal cell, etc.. The results showed that stoma exists only on the lower epidermis and its distribution is irregular, and leaf epidermis consist of epidermis cells, stoma complexes and bushy trichomes including glandular hair and non-glandular hair. On the upper epidermis, anticlinal wall caves in sinuous groove to countercheck the transpiration. Evidences from leaf morphological structures serve as another proof on drought-resistant mechanisms. Some strumaes distributing regularly are hypothesized as oxalic calcium on the lower epidermis under laser scanning confocal microscopy （LSCM） with Fluo-3/AM, which can increase their endurance to drought stress. Therefore, the above characteristics of Flos Lonicerae can reduce the loss of water and make Japanese honeysuckle and Wild Honeysuckle adapt to the droughty environment at Karst area in southwest China. However, there is some difference of the two species. From the SEM （Scanning Electron Microscopy） result, it is shown that on the upper epidermis, some glandular hair regularly present along the midrib of Japanese honeysuckle, but Wild Honeysuckle has no glandular hair on the upper epidermis, which can verify the relationships of Flos Lonicerae species and provide the significance for classification of Flos Lonicerae.展开更多
在法律判决预测领域中,案件描述通常具有相似的结构,而现有的预测方法容易忽略不同案件间的要素差异,难以有效利用案件要素特征,导致模型预测准确率不高.此外,罪名预测任务还面临易混淆罪名问题.针对上述问题,提出了一种融合案件要素和...在法律判决预测领域中,案件描述通常具有相似的结构,而现有的预测方法容易忽略不同案件间的要素差异,难以有效利用案件要素特征,导致模型预测准确率不高.此外,罪名预测任务还面临易混淆罪名问题.针对上述问题,提出了一种融合案件要素和案件属性的罪名预测多任务学习模型(Case Elements And Attributes Multi-Task Learning Model,简称CEAT-MLM),通过挖掘案件要素及案件属性与罪名之间的关联关系,将案件属性预测和罪名预测进行联合建模,达到提升罪名预测准确率的目标.实验结果表明,本文提出的模型相较于通用的文本分类模型具有显著的性能提升,并与法律判决领域的典型模型相比,Macro-F1得分提升了1.76%.展开更多
文摘Evaluating the adversarial robustness of classification algorithms in machine learning is a crucial domain.However,current methods lack measurable and interpretable metrics.To address this issue,this paper introduces a visual evaluation index named confidence centroid skewing quadrilateral,which is based on a classification confidence-based confusion matrix,offering a quantitative and visual comparison of the adversarial robustness among different classification algorithms,and enhances intuitiveness and interpretability of attack impacts.We first conduct a validity test and sensitive analysis of the method.Then,prove its effectiveness through the experiments of five classification algorithms including artificial neural network(ANN),logistic regression(LR),support vector machine(SVM),convolutional neural network(CNN)and transformer against three adversarial attacks such as fast gradient sign method(FGSM),DeepFool,and projected gradient descent(PGD)attack.
基金supported by the NationalNatural Science Foundation of China Nos.62302167,U23A20343Shanghai Sailing Program(23YF1410500)Chenguang Program of Shanghai Education Development Foundation and Shanghai Municipal Education Commission(23CGA34).
文摘Confusing object detection(COD),such as glass,mirrors,and camouflaged objects,represents a burgeoning visual detection task centered on pinpointing and distinguishing concealed targets within intricate backgrounds,leveraging deep learning methodologies.Despite garnering increasing attention in computer vision,the focus of most existing works leans toward formulating task-specific solutions rather than delving into in-depth analyses of methodological structures.As of now,there is a notable absence of a comprehensive systematic review that focuses on recently proposed deep learning-based models for these specific tasks.To fill this gap,our study presents a pioneering review that covers both themodels and the publicly available benchmark datasets,while also identifying potential directions for future research in this field.The current dataset primarily focuses on single confusing object detection at the image level,with some studies extending to video-level data.We conduct an in-depth analysis of deep learning architectures,revealing that the current state-of-the-art(SOTA)COD methods demonstrate promising performance in single object detection.We also compile and provide detailed descriptions ofwidely used datasets relevant to these detection tasks.Our endeavor extends to discussing the limitations observed in current methodologies,alongside proposed solutions aimed at enhancing detection accuracy.Additionally,we deliberate on relevant applications and outline future research trajectories,aiming to catalyze advancements in the field of glass,mirror,and camouflaged object detection.
文摘Recently,in Northeast Asian situation there have occurred escalation of tension and intensification of antagonism seldom seen before.At the same time,however,there are also explorations and signs of restarting dialogue and cooperation that have stagnated since long ago.Development trends of the situation are confusing to the eye,but reconciliation and cooperation remain the way-out for
文摘Lactic acid bacteria(LAB)were probiotics that produced a variety of metabolites.As a probiotic,exopolysaccharide(EPS)from LAB was widely used in the fields of food and medicine.Morphological observation,API 50 CHL test and 16S rDNA were used to identify the species characteristics of EPS producing LAB,and to explore the strain tolerance characteristics.The results showed that the strain P2 was Weissella confusa and named W.confusa P2.The survival rate of this strain was more than 60%in the range of 0~0.25 g·L^(-1) bile salt concentration,and it could survive well at pH 4.0~7.5 and NaCl concentration 0~4 g·L^(-1).Moreover,the EPS content of W.confusa P2 reached a maximum of 16.06±0.33 g·L^(-1) when the fermentation time was 36 h.W.confusa P2 may be a potential candidate strain in the field of food fermentation and possibly probiotics.Therefore,as a strain with tolerance and high EPS yield,W.confusa P2 not only had significant application potential in the food industry,but also provided new resources and directions for the development of probiotic preparations.
文摘The term“nipple confusion”accurately describes the confusion newborns experience between their mother’s nipple and an artificial nipple during feeding.Specifically,it refers to the feeding habits infants develop based on their initial feeding experiences after birth.Infants accustomed to the maternal nipple often resist bottle-feeding;conversely,those accustomed to bottle-feeding may reject the maternal nipple.This confusion is particularly common among infants receiving mixed feeding.
文摘Nipple confusion,a term that accurately describes the confusion that occurs between the mother’s nipple and the artificial teat during feeding in newborns.Specifically,it refers to the fact that babies develop specific breastfeeding habits after birth,based on their initial feeding experience.For babies who are accustomed to their mother’s nipple,they tend to show resistance to bottle feeding;on the contrary,those who have adapted to bottle feeding may refuse to accept their mother’s nipple.This confusion is particularly common among mixed-feeding babies.
基金This study was supported by the Ministry of Sciences and Technology of China (No.2005DIB3J067)the National Science Foundation of China (No.40572107, No.40231008, No.40672165 and No.30600074)+2 种基金the Chongqing Science & Technology Commission (No.2005AB7006)the Open Fund and Key Subject of Physical Geog-raphy, Southwest Normal University of China (No.250-411110)the Open Fund of Key Laboratory of Chinese Academy of Geological Sci-ences (No.KL05-20).
文摘The leaf epidermis of Japanese honeysuckle （Lonicera japonica Thunb.） and Wild Honeysuckle （Lonicera confusa） in the genus of Flos Lonicerae were mainly observed by scanning electron microscopes （SEM） to study the characteristics of stomata, trichomes and dermal cell, etc.. The results showed that stoma exists only on the lower epidermis and its distribution is irregular, and leaf epidermis consist of epidermis cells, stoma complexes and bushy trichomes including glandular hair and non-glandular hair. On the upper epidermis, anticlinal wall caves in sinuous groove to countercheck the transpiration. Evidences from leaf morphological structures serve as another proof on drought-resistant mechanisms. Some strumaes distributing regularly are hypothesized as oxalic calcium on the lower epidermis under laser scanning confocal microscopy （LSCM） with Fluo-3/AM, which can increase their endurance to drought stress. Therefore, the above characteristics of Flos Lonicerae can reduce the loss of water and make Japanese honeysuckle and Wild Honeysuckle adapt to the droughty environment at Karst area in southwest China. However, there is some difference of the two species. From the SEM （Scanning Electron Microscopy） result, it is shown that on the upper epidermis, some glandular hair regularly present along the midrib of Japanese honeysuckle, but Wild Honeysuckle has no glandular hair on the upper epidermis, which can verify the relationships of Flos Lonicerae species and provide the significance for classification of Flos Lonicerae.
文摘在法律判决预测领域中,案件描述通常具有相似的结构,而现有的预测方法容易忽略不同案件间的要素差异,难以有效利用案件要素特征,导致模型预测准确率不高.此外,罪名预测任务还面临易混淆罪名问题.针对上述问题,提出了一种融合案件要素和案件属性的罪名预测多任务学习模型(Case Elements And Attributes Multi-Task Learning Model,简称CEAT-MLM),通过挖掘案件要素及案件属性与罪名之间的关联关系,将案件属性预测和罪名预测进行联合建模,达到提升罪名预测准确率的目标.实验结果表明,本文提出的模型相较于通用的文本分类模型具有显著的性能提升,并与法律判决领域的典型模型相比,Macro-F1得分提升了1.76%.
