Abstract:In order to investigate structure-property relationships, the catalytic properties of gold nanoparticles were evaluated in the reduction of 4-nitrophenol by NaBH 4 . Using catalyst suspensions with identical amounts of gold, the following key results were obtained: first, the nanostars showed increased activity as compared to spherical gold nanoparticles; second, larger gold nanostars showed higher activity, likely because of the abundance of flat/spiky features on these particles, which show high metal utilization; third, treatment of the nanostar colloid with cucurbit [7]uril can be used to balance catalyst stability and activity; fourth, as expected from the decreasing surface atom fraction, the specific activity of the spherical nanoparticles decreased with increasing particle size.

Although macrocyclic host engineered nanoparticles have drawn a lot of attention in the past decade, little attention was paid to the binding affinity between the host and guest assembled on nanoparticles. In this work, the interactions between cucurbit[7]­uril (CB[7]) and three guests with phenyl (1), adamantyl (2), and ferrocene (3) groups on a gold surface were investigated. Their interactions were classified into four groups. The binding constants between CB[7] and three guests in the solution (group I) were determined by proton nuclear magnetic resonance (1H NMR) and isothermal titration calorimetry (ITC) experiments. The binding constants between CB[7]-stabilized gold nanosphere and guests 1 and 3 (group II) and those between 1-stabilized gold nanosphere and CB[7] (group III) were determined by ITC experiments. The binding constants determined in groups II and III decrease to approximately 20% of those in group I. The interactions between CB[7] and three guests on gold (group IV) were measured by single-molecule force spectroscopy (SMFS). Much lower binding constants than those in other groups were deduced. The hydrophobic and ion–dipole interactions play important roles in the pulling-out process in a thorough SMFS analysis. Furthermore, the CB[7]-capped gold nanospheres with different sizes were prepared, and their binding constant with the free guest (group II) depends on the nanoparticle size. All of these results are of great benefit to the rational design of macrocyclic host engineered nanoparticles for drug delivery and many other applications.

The data imbalance problem is a crucial issue for the multi-label text classification. Some existing works tackle it by proposing imbalanced loss objectives instead of the vanilla cross-entropy loss, but their performances remain limited in the cases of extremely imbalanced data. We propose a hybrid solution which adapts general networks for the head categories, and few-shot techniques for the tail categories. We propose a Hybrid-Siamese Convolutional Neural Network (HSCNN) with additional technical attributes, i.e., a multi-task architecture based on Single and Siamese networks; a category-specific similarity in the Siamese structure; a specific sampling method for training HSCNN. The results using two benchmark datasets and three loss objectives show that our method can improve the performance of Single networks with diverse loss objectives on the tail or entire categories.

No abstract

No abstract