In this study, we propose an incremental learning approach based on a machine-machine interaction via relative attribute feedbacks that exploit comparative relationships among top level image categories. One machine a...In this study, we propose an incremental learning approach based on a machine-machine interaction via relative attribute feedbacks that exploit comparative relationships among top level image categories. One machine acts as ＇Student （S）＇ with initially limited information and it endeavors to capture the task domain gradually by questioning its mentor on a pool of unlabeled data. The other machine is ＇Teacher （T）＇ with the implicit knowledge for helping S on learning the class models. T initiates relative attributes as a communication channel by randomly sorting the classes on attribute space in an unsupervised manner. S starts modeling the categories in this intermediate level by using only a limited number of labeled data. Thereafter, it first selects an entropy-based sample from the pool of unlabeled data and triggers the conversation by propagating the selected image with its belief class in a query. Since T already knows the ground truth labels, it not only decides whether the belief is true or false, but it also provides an attribute-based feedback to S in each case without revealing the true label of the query sample if the belief is false. So the number of training data is increased virtually by dropping the falsely predicted sample back into the unlabeled pool. Next, S updates the attribute space which, in fact, has an impact on T＇s future responses, and then the category models are updated concurrently for the next run. We experience the weakly supervised algorithm on the real world datasets of faces and natural scenes in comparison with direct attribute prediction and semi-supervised learning approaches, and a noteworthy performance increase is achieved.展开更多
The effects of temperature and step-down relative humidity controlled hot-air drying(THC-HAD)on the drying kinetics,energy efficiency and quality,i.e.,rehydration ratio(RR),color parameters(L*,a*,b*),total color diffe...The effects of temperature and step-down relative humidity controlled hot-air drying(THC-HAD)on the drying kinetics,energy efficiency and quality,i.e.,rehydration ratio(RR),color parameters(L*,a*,b*),total color difference(ΔE*),Panax notoginseng saponins(PNS)content,and ginsenosides content(R1,Rg1,Re,Rd,Rb1)of Panax notoginseng roots were evaluated.The drying time was significantly affected by the drying temperature followed by the relative humidity(RH)of the drying air.Special combination of drying conditions,i.e.,drying temperature of 50°C,relative humidity of 40%for 3 h and then continuous dehumidification from 40%to 8%allowed to shorten the drying time by 25%compared to drying at the same temperature and continuous dehumidification.The longer was the drying time under constant high RH of drying air,the lower was the RR of dried samples.The step-down RH strategy contributed to the formation of a porous structure,enhancement of drying efficiency and quality improvement.Generally,the ginsenosides content increased with the increase in temperature,while no obvious trend was recorded for ginsenoside R1.The contents of the ginsenoside R1,Rg1,Rb1 and PNS decreased with the increase in the drying time under constant high RH.Taking into account the drying time,energy consumption and quality attributes,drying at the temperature of 50°C,constant RH of 40%for 3 h and then step-down RH from 40%to 8%was proposed as the most favorable combination of drying conditions for dehydration of whole Panax notoginseng roots.展开更多
The thawing-melting of the permafrost damages the subground of highways on the Qinghai-Tibet Plateau.With the application of ground-penetrating-radar(GPR)technology,the maximum permafrost melting interface can be effe...The thawing-melting of the permafrost damages the subground of highways on the Qinghai-Tibet Plateau.With the application of ground-penetrating-radar(GPR)technology,the maximum permafrost melting interface can be effectively distinctly differentiated and imaged.A hierarchical feature of the permafrost region is shown clearly on the imaging profile of GPR data.The complete ablation zone or part of it is displayed distinctly.In addition,the details of subsurface layers can be effectively characterized by GPR attribute-analysis technology.With the attribute calculation and filter,the instantaneous amplitude,instantaneous frequency,and relative wave impedance can be applied in a more efficient way to divide the complete ablation zone,part of the ablation and non-ablation interface.The relative distribution of water content in a seasonally thawing permafrost region can be obtained through a comprehensive GPR attribute analysis.展开更多
The clustering on categorical variables has received intensive attention. In dataset with categorical features, some features show the superior performance on clustering procedure. In this paper, we propose a simple m...The clustering on categorical variables has received intensive attention. In dataset with categorical features, some features show the superior performance on clustering procedure. In this paper, we propose a simple method to find such distinctive features by comparing pooled within-cluster mean relative difference and then partition the data upon such features and give subspace of the subgroups. The applications on zoo data and soybean data illustrate the performance of the proposed method.展开更多
A new algorithm based on rough core was proposed to extract all relative-attribute reducts in decision information systems of large-scale records. In the algorithm, the rough core of the decision-making information sy...A new algorithm based on rough core was proposed to extract all relative-attribute reducts in decision information systems of large-scale records. In the algorithm, the rough core of the decision-making information system is first calculated. Then, an approach based on a top-down strategy is adopted to select the non-core condition attributes and generate candidate relative-attribute reducts. Finally, the set of all relative-attribute reducts is obtained by pruning the candidate relative-attribute reducts. Experimental results show that the proposed algorithm is superior to the other methods such as the algorithm without computing core, the exhaustive method and the discernibility matrix method in extracting all relative-attribute reducts for large-scale data sets.展开更多
文摘In this study, we propose an incremental learning approach based on a machine-machine interaction via relative attribute feedbacks that exploit comparative relationships among top level image categories. One machine acts as ＇Student （S）＇ with initially limited information and it endeavors to capture the task domain gradually by questioning its mentor on a pool of unlabeled data. The other machine is ＇Teacher （T）＇ with the implicit knowledge for helping S on learning the class models. T initiates relative attributes as a communication channel by randomly sorting the classes on attribute space in an unsupervised manner. S starts modeling the categories in this intermediate level by using only a limited number of labeled data. Thereafter, it first selects an entropy-based sample from the pool of unlabeled data and triggers the conversation by propagating the selected image with its belief class in a query. Since T already knows the ground truth labels, it not only decides whether the belief is true or false, but it also provides an attribute-based feedback to S in each case without revealing the true label of the query sample if the belief is false. So the number of training data is increased virtually by dropping the falsely predicted sample back into the unlabeled pool. Next, S updates the attribute space which, in fact, has an impact on T＇s future responses, and then the category models are updated concurrently for the next run. We experience the weakly supervised algorithm on the real world datasets of faces and natural scenes in comparison with direct attribute prediction and semi-supervised learning approaches, and a noteworthy performance increase is achieved.
基金supported in part by the Hebei Province Key Research and Development Project(Grant No.203777119D,19227210D)in part by the Scientific Research Projects of Universities in Hebei Province(Grant No.ZD2021056)in part by the Hebei Province College and Middle School Students Science and Technology Innovation Ability Cultivation Project(Grant No.2021H060505)and part by China Agriculture Research System of MOF and MARA(CARS-21).
文摘The effects of temperature and step-down relative humidity controlled hot-air drying(THC-HAD)on the drying kinetics,energy efficiency and quality,i.e.,rehydration ratio(RR),color parameters(L*,a*,b*),total color difference(ΔE*),Panax notoginseng saponins(PNS)content,and ginsenosides content(R1,Rg1,Re,Rd,Rb1)of Panax notoginseng roots were evaluated.The drying time was significantly affected by the drying temperature followed by the relative humidity(RH)of the drying air.Special combination of drying conditions,i.e.,drying temperature of 50°C,relative humidity of 40%for 3 h and then continuous dehumidification from 40%to 8%allowed to shorten the drying time by 25%compared to drying at the same temperature and continuous dehumidification.The longer was the drying time under constant high RH of drying air,the lower was the RR of dried samples.The step-down RH strategy contributed to the formation of a porous structure,enhancement of drying efficiency and quality improvement.Generally,the ginsenosides content increased with the increase in temperature,while no obvious trend was recorded for ginsenoside R1.The contents of the ginsenoside R1,Rg1,Rb1 and PNS decreased with the increase in the drying time under constant high RH.Taking into account the drying time,energy consumption and quality attributes,drying at the temperature of 50°C,constant RH of 40%for 3 h and then step-down RH from 40%to 8%was proposed as the most favorable combination of drying conditions for dehydration of whole Panax notoginseng roots.
基金supported by the Fundamental Research Funds for the Central Universities (2015JBM064)the 49th Project Sponsored by the Scientific Research Foundation for the Returned Overseas Chinese Scholars infrastructure in State Education Ministry,and the research project entitled "The freezing injury evaluation of subgrade and remediation technology research in Shenchi-Shuozhou Railway" (No. 2015-10)
文摘The thawing-melting of the permafrost damages the subground of highways on the Qinghai-Tibet Plateau.With the application of ground-penetrating-radar(GPR)technology,the maximum permafrost melting interface can be effectively distinctly differentiated and imaged.A hierarchical feature of the permafrost region is shown clearly on the imaging profile of GPR data.The complete ablation zone or part of it is displayed distinctly.In addition,the details of subsurface layers can be effectively characterized by GPR attribute-analysis technology.With the attribute calculation and filter,the instantaneous amplitude,instantaneous frequency,and relative wave impedance can be applied in a more efficient way to divide the complete ablation zone,part of the ablation and non-ablation interface.The relative distribution of water content in a seasonally thawing permafrost region can be obtained through a comprehensive GPR attribute analysis.
文摘The clustering on categorical variables has received intensive attention. In dataset with categorical features, some features show the superior performance on clustering procedure. In this paper, we propose a simple method to find such distinctive features by comparing pooled within-cluster mean relative difference and then partition the data upon such features and give subspace of the subgroups. The applications on zoo data and soybean data illustrate the performance of the proposed method.
文摘A new algorithm based on rough core was proposed to extract all relative-attribute reducts in decision information systems of large-scale records. In the algorithm, the rough core of the decision-making information system is first calculated. Then, an approach based on a top-down strategy is adopted to select the non-core condition attributes and generate candidate relative-attribute reducts. Finally, the set of all relative-attribute reducts is obtained by pruning the candidate relative-attribute reducts. Experimental results show that the proposed algorithm is superior to the other methods such as the algorithm without computing core, the exhaustive method and the discernibility matrix method in extracting all relative-attribute reducts for large-scale data sets.
