TaijiGNN: A New Cycle-Consistent Generative Neural Network for High-Quality Bidirectional Transformation between RGB and Multispectral Domains.

Since multispectral images (MSIs) and RGB images (RGBs) have significantly different definitions and severely imbalanced information entropies, the spectrum transformation between them, especially reconstructing MSIs from RGBs, is a big challenge. We propose a new approach, the Taiji Generative Neural Network (TaijiGNN), to address the above-mentioned problems. TaijiGNN consists of two generators, G_MSI, and G_RGB. These two generators establish two cycles by connecting one generator's output with the other's input. One cycle translates the RGBs into the MSIs and converts the MSIs back to the RGBs. The other cycle does the reverse. The cycles can turn the problem of comparing two different domain images into comparing the same domain images. In the same domain, there are neither different domain definition problems nor severely underconstrained challenges, such as reconstructing MSIs from RGBs. Moreover, according to several investigations and validations, we effectively designed a multilayer perceptron neural network (MLP) to substitute the convolutional neural network (CNN) when implementing the generators to make them simple and high performance. Furthermore, we cut off the two traditional CycleGAN's identity losses to fit the spectral image translation. We also added two consistent losses of comparing paired images to improve the two generators' training effectiveness. In addition, during the training process, similar to the ancient Chinese philosophy Taiji's polarity Yang and polarity Yin, the two generators update their neural network parameters by interacting with and complementing each other until they all converge and the system reaches a dynamic balance. Furthermore, several qualitative and quantitative experiments were conducted on the two classical datasets, CAVE and ICVL, to evaluate the performance of our proposed approach. Promising results were obtained with a well-designed simplistic MLP requiring a minimal amount of training data. Specifically, in the CAVE dataset, to achieve comparable state-of-the-art results, we only need half of the dataset for training; for the ICVL dataset, we used only one-fifth of the dataset to train the model, but obtained state-of-the-art results.

Multi-matrices factorization with application to missing sensor data imputation.

We formulate a multi-matrices factorization model (MMF) for the missing sensor data estimation problem. The estimation problem is adequately transformed into a matrix completion one. With MMF, an n-by-t real matrix, R, is adopted to represent the data collected by mobile sensors from n areas at the time, T1, T2, ..., Tt, where the entry, Rij, is the aggregate value of the data collected in the ith area at Tj. We propose to approximate R by seeking a family of d-by-n probabilistic spatial feature matrices, U(1), U(2), ..., U(t), and a probabilistic temporal feature matrix, [formula in text]. We also present a solution algorithm to the proposed model. We evaluate MMF with synthetic data and a real-world sensor dataset extensively. Experimental results demonstrate that our approach outperforms the state-of-the-art comparison algorithms.

Molten salt electrolytic synthesis of porous carbon from SiC and its application in supercapacitors.

Nanoscale porous carbide-derived carbon (CDC) microspheres were successfully synthesized via the electrolysis etching of nano-SiC microsphere powder precursors with a particle diameter of 200 to 500 nm in molten CaCl2. Electrolysis was conducted at 900 °C for 14 h in argon at an applied constant voltage of 3.2 V. The results show that the obtained product is SiC-CDC, which is a mixture of amorphous carbon and a small quantity of ordered graphite with a low degree of graphitization. Similar to the SiC microspheres, the obtained product retained its original shape. The specific surface area was 734.68 m2 g-1. The specific capacitance of the SiC-CDC was 169 F g-1, and it exhibited excellent cycling stability (98.01% retention of the initial capacitance after 5000 cycles) at a current density of 1000 mA g-1.

[Microbial Diversity and Influencing Factors in a Small Watershed in Winter].

The water quality of a small watershed will directly affect the water quality safety of the entire watershed. The Lidu small watershed of the Fulin District in Chongqing was selected for this study. The characteristics of runoff and the microbial community in the receiving water body were analyzed by high-throughput sequencing technology. The effects of the physical and chemical factors of the water body on the microbial diversity in the water body were also investigated. The results showed that the small watershed of Lidu was an important source of nitrite and dissolved organic carbon for the receiving water. There was no significant difference in the microbial richness and diversity of the water flowing through different pollution sources in the small watershed, and the microbial diversity index was negatively correlated with the total phosphorus (r -0.79--0.84, P ≤ 0.01) and phosphate (r -0.71--0.80, P ≤ 0.01) of the water. At the phylum level, Actinobacteria (37.33% ±14.69%) accounted for the largest proportion of runoff flowing through the agricultural area, followed by Proteobacteria (32.53% ±7.18%), Cyanobacteria (14.65% ±22.64%), Bacteroidetes (8.50% ±3.67%), and others. Proteobacteria (43.67% ±10.80%) accounted for the largest proportion of runoff flowing through the industrial area, followed by Bacteroidetes (25.33% ±11.97%), Actinobacteria (24.17% ±14.66%), Firmicutes (2.53% ±0.72%), and others. At the genus level, hgcI_clade (19.08% ±13.46%) accounted for the largest proportion of runoff flowing through the agricultural area, followed by CL 500-29 _marine_group (7.40% ±4.44%), Limnohabitans (7.05% ±3.14%), and others. Flavobacterium (20.40% ±12.37%) accounted for the largest proportion of runoff flowing through the industrial area, followed by hgcI_clade (15.30% ±11.11%), Acinetobacter (9.33% ±11.50%), and others. The Flavobacterium in the runoff water may be related to the input of industrial sources in the watershed. Nitrous nitrogen, total phosphorus, phosphate, pH, and Zn2+ were the main environmental factors that affected the microbial community in the small watershed. They can explain the variance of microbial community 0.26, 0.21, 0.20, 0.15, and 0.14, respectively. DOC, nitrate nitrogen, and heavy metal ions such as Cu2+, Zn2+, Mn2+, and As3+ were also related to the proportion of some microorganisms in the water.

Hierarchical network modeling with multidimensional information for aquatic safety management in the cold chain.

The cold-chain information has characterized by the loss and dispersion according to the different collecting methods. The description for the quality decay factors of aquatic products can be defined as the multidimensional information. A series of nodes with multidimensional information are assembled to be hierarchies aiming at describing the environment conditions and locations in the supply chain. Each of the single hierarchy levels constitutes a sequence of node information in a network, which is applied as internal information analysis. The cross-layer information structure is defined as "bridge" information which is able to record the information transmissions among every hierarchy from the point of view of the whole chain. The study has established a novel structured modeling to describe the cold chain of aquatic products based on a network-hierarchy framework. An organized and sustainable transmission process can be built and recorded by the multidimensional attributes for the whole course of cold chain of aquatic products. In addition, seamless connections among every hierarchy are attainable by the environmental information records continuously to monitor the quality of aquatic products. The quality assessments and shelf life predictions are estimated properly as the risk control in order to monitor and trace the safety of aquatic products under the supply chain perspective.