Search

Recently, a new method of transfer learning called adversarial domain adaptation, embeds the idea of the generative adversarial networks (GAN) into the deep networks. It can learn the transferable representation of data for domain adaptation by the thought of the GAN. Although this method can extract the common features of the source domain data and target domain data, and effectively transfer knowledge between different domains, the existing adversarial domain adaptation algorithms cannot effectively retain the local features of the target domain. However, some features of the target domain data may significantly improve the classification accuracy. In order to avoid the destruction of the local features of the original data due to adversarial learning, a multi-task neural network is used to retain the local features of the target domain data. A model of deep adversarial-reconstruction-classification networks (DARCN) is proposed. DARCN is inspired by the auto-encoder. On the basis of adversarial domain adap-tation, the decoding part of the auto-encoder is added, which can effectively reconstruct the original data from low-dimensional features. The model learns shared coding representations for the following tasks: supervised classification of labeled source domain data, unsupervised reconstruction of unlabeled target domain data and indistinguishability of source domain and target domain. Finally, the classification loss of the label classifier and the reconstruction loss of the decoder are minimized, and the classification loss of the domain discriminator is maximized. The gradient descent method can effectively solve such optimization problems. The experimental results prove that the preservation of local features of target domain is critical for domain adaptation tasks.

This work investigated the underlying mechanism of thermo-ultrasonic treatment to improve the nutritional properties and emulsion stability of sea bass fish head broth. The effects of ultrasonication on the processing of fish broth were compared with boiling water treatment. The nutritional properties of fish broth mainly include protein, fat, total sugar, 5′-nucleotide and free amino acid content. To achieve a similar effect of nutrient extraction, the thermo-ultrasonic treatment required a shorter time (30 min) than boiling water (120 min). The water-soluble protein, fat and total sugar contents were at their maximum at 120 min of the thermo-ultrasonic treatment. In particular, the fat content increased with the time of thermo-ultrasonic treatment from 0.58% to 2.70%. The emulsion structure of the fish soup was characterized by measuring its color and particle size, using optical microscopy and confocal laser scanning microscopy, and determining its storage stability. Thermo-ultrasonic treatment reduced the particle size of the fish broth emulsion and the fat globules became smaller and more homogeneous. Ultrasonication not only accelerated the nutritional and flavor content of the fish head broth, but also reduced the particle size and enhanced the stability of the emulsified system of the fish broth. The fish head tissue was more severely disrupted by the cavitation effect of an ultrasound, and nutrients migrated more and faster. This was mainly due to the cavitation and mechanical breaking force of the ultrasound on the fish head tissue and the fat globules of the fish broth. Altogether, these findings suggest that the thermo-ultrasonic treatment technique is useful for processing nutrient-rich, storage-stable and ready-to-eat fish head broth. [ABSTRACT FROM AUTHOR]

Objective To analyze the incidence trend and epidemiological characteristics of dengue fever in Fujian Province from 2009 to 2023, identify the high-incidence population and hotspot areas, and a scientific basis for formulating targeted prevention and control measures. Methods The surveillance data of dengue fever during 2009-2023 in Fujian Province were obtained from the National Disease Reporting Information System and analyzed using descriptive epidemiological methods, and the spatial autocorrelation analysis of dengue fever incidence at county/district levels was performed with ArcGlS 10.8. Results A total of 3 586 cases of dengue fever were reported in Fujian Province from 2009 to 2023, with an average annual incidence rate of 0.61/100 000, showing an overall increasing trend year by year (Z=18.35, P< 0.001). Among them, there were 2 362 local cases and 1 224 imported cases, with a local/imported ratio of 1.93:1 (2 362/1 224). The local cases were mainly distributed in Fuzhou, Putian, and Nanping, accounting for 87.81% (2 074/2 362). The imported cases were mainly distributed in Quanzhou, Fuzhou, and Xiamen, accounting for 72.55% (888/1 224). The local/imported ratio in Fuzhou, Nanping, and Putian was 6.20:1 (1 557/251), 3.92:1 (145/37), and 3.32:1 (372/112), respectively, all significantly higher than the provincial average level. The spatial distribution of the local cases showed a clustered pattern, with 12 hotspots of incidence, all located in the counties and districts under the jurisdiction of Fuzhou. The onset of the disease had obvious seasonality: imported cases from other countries or regions mainly occurred from July to September (accounting for 45.73%, 519/1 135), imported cases from other provinces mainly occurred from August to November (accounting for 93.25%, 83/89), and local cases mainly occurred from August to October (accounting for 97.50%, 2 303/2 362). The gender ratio for local cases was 0.88:1, and for imported cases, it was 3.04:1. The average age of onset for local cases was 48 years old (Q1=32 years old, Q3=62 years old), while the average age for imported cases was 35 years old (Q1=28 years old, Q3=44 years old). Conclusions From 2009 to 2023, the dengue fever epidemic in Fujian Province showed a continuous upward trend, with spatiotemporal clustering of incidence. Therefore, prevention and control measures should be strengthened in key areas and populations during the high-incidence season. [ABSTRACT FROM AUTHOR]

Hybrid metal halides are an extraordinary class of optoelectronic materials with extensive applications. To further diversify and study the in‐depth structure‐property relations, we report here a new family of 21 zero‐dimensional hybrid bimetallic chlorides with the general formula A(L)n[BClm] (A=rare earth (RE), alkaline earth metals and Mn; L=solvent ligand; and B=Sb, Bi and Te). The RE(DMSO)8[BCl6] (RE=La, Ce, Sm, Eu, Tb, and Dy; DMSO=dimethyl sulfoxide) series shows broadband emission attributed to triplet radiative recombination from Sb and Bi, incorporating the characteristic emission of RE metals, where Eu(DMSO)8[BiCl6] shows a staggering PL quantum yield of 94 %. The pseudo‐octahedral [SbCl5] with Cl vacancy in AII(DMSO)6[SbCl5] (AII=Mg, Ca and Mn) and the square pyramidal [SbCl5] in AII(TMSO)6[SbCl5] (TMSO=tetramethylene sulfoxide) enhance the stereoactive expression of the 5 s2 lone pairs of Sb3+, giving rise to the observation of dual‐band emission of singlet and triplet emission, respectively. A series of Te(IV) analogues have been characterized, showing blue‐light‐excitable single‐band emission. This work expands the materials space for hybrid bimetallic halides with an emphasis on harnessing the RE elements, and provides important insights into designing new emitters and regulating their properties. [ABSTRACT FROM AUTHOR]

In this study, phosphoric acid activation was employed to synthesize nitrogendoped mesoporous activated carbon (designated as MR1) from Lentinus edodes (shiitake mushroom) residue, while aiming to efficiently remove acetaminophen (APAP), carbamazepine (CBZ), and metronidazole (MNZ) from aqueous solutions. We characterized the physicochemical properties of the produced adsorbents using scanning electron microscopy (SEM), nitrogen adsorption isotherms, and X-ray photoelectron spectroscopy (XPS). MR1, MR2, and MR3 were prepared using phosphoric acid impregnation ratios of 1, 2, and 3 mL/g, respectively. Notably, MR1 exhibited a significant mesoporous structure with a volume of 0.825 cm3/g and a quaternary nitrogen content of 2.6%. This endowed MR1 with a high adsorption capacity for APAP, CBZ, and MNZ, positioning it as a promising candidate for water purification applications. The adsorption behavior of the contaminants followed the Freundlich isotherm model, suggesting a multilayer adsorption process. Notably, MR1 showed excellent durability and recyclability, maintaining 95% of its initial adsorption efficiency after five regeneration cycles and indicating its potential for sustainable use in water treatment processes. [ABSTRACT FROM AUTHOR]