The Journal of Supercomputing - With the development of cloud computing application, attribute-based encryption (ABE) with flexibly fine-grained data access control is widely adopted. However,... 相似文献
The Journal of Supercomputing - In location-based social network platforms, the point-of-interest(POI) recommendation is an essential function to serve users. The existing POI recommendation... 相似文献
The heavy reliance on data is one of the major reasons that currently limit the development of deep learning. Data quality directly dominates the effect of deep learning models, and the long-tailed distribution is one of the factors affecting data quality. The long-tailed phenomenon is prevalent due to the prevalence of power law in nature. In this case, the performance of deep learning models is often dominated by the head classes while the learning of the tail classes is severely underdeveloped. In order to learn adequately for all classes, many researchers have studied and preliminarily addressed the long-tailed problem. In this survey, we focus on the problems caused by long-tailed data distribution, sort out the representative long-tailed visual recognition datasets and summarize some mainstream long-tailed studies. Specifically, we summarize these studies into ten categories from the perspective of representation learning, and outline the highlights and limitations of each category. Besides, we have studied four quantitative metrics for evaluating the imbalance, and suggest using the Gini coefficient to evaluate the long-tailedness of a dataset. Based on the Gini coefficient, we quantitatively study 20 widely-used and large-scale visual datasets proposed in the last decade, and find that the long-tailed phenomenon is widespread and has not been fully studied. Finally, we provide several future directions for the development of long-tailed learning to provide more ideas for readers.
Inspired by mussel‐adhesion phenomena in nature, polydopamine (PDA) coatings are a promising route to multifunctional platforms for decorating various materials. The typical self‐polymerization process of dopamine is time‐consuming and the coatings of PDA are not reusable. Herein, a reusable and time‐saving strategy for the electrochemical polymerization of dopamine (EPD) is reported. The PDA layer is deposited on vertically aligned TiO2 nanotube arrays (NTAs). Owing to the abundant catechol and amine groups in the PDA layer, uniform Pt nanoparticles (NPs) are deposited onto the TiO2 NTAs and can effectively prevent the recombination of electron–hole pairs generated from photo‐electrocatalysis and transfer the captured electrons to participate in the photo‐electrocatalytic reaction process. Compared with pristine TiO2 NTAs, the as‐prepared Pt@TiO2 NTA composites exhibit surface‐enhanced Raman scattering sensitivity for detecting rhodamine 6G and display excellent UV‐assisted self‐cleaning ability, and also show promise as a nonenzymatic glucose biosensor. Furthermore, the mussel‐inspired electropolymerization strategy and the fast EPD‐reduced nanoparticle decorating process presented herein can be readily extended to various functional substrates, such as conductive glass, metallic oxides, and semiconductors. It is the adaptation of the established PDA system for a selective, robust, and generalizable sensing system that is the emphasis of this work. 相似文献
Optical fluorescence imaging is an important strategy to explore the mechanism of virus–host interaction. However, current fluorescent tag labeling strategies often dampen viral infectivity. The present study explores an in situ fluorescent labeling strategy in order to preserve viral infectivity and precisely monitor viral infection in vivo. In contrast to pre‐labeling strategy, mice are first intranasally infected with azide‐modified H5N1 pseudotype virus (N3‐H5N1p), followed by injection of dibenzocyclooctyl (DBCO)‐functionalized fluorescence 6 h later. The results show that DBCO dye directly conjugated to N3‐H5N1p in lung tissues through in vivo bioorthogonal chemistry with high specificity and efficacy. More remarkably, in situ labeling rather than conventional prelabeling strategy effectively preserves viral infectivity and immunogenicity both in vitro and in vivo. Hence, in situ bioorthogonal viral labeling is a promising and reliable strategy for imaging and tracking viral infection in vivo. 相似文献