Search

Atmospheric CO2 concentration is increasing, largely due to anthropogenic activities. Previous studies of individual free-air CO2 enrichment (FACE) experimental sites have shown significant impacts of elevated CO2 (eCO2) on soil microbial communities; however, no common microbial response patterns have yet emerged, challenging our ability to predict ecosystem functioning and sustainability in the future eCO2 environment. Here we analyzed 66 soil microbial communities from five FACE sites, and showed common microbial response patterns to eCO2, especially for key functional genes involved in carbon and nitrogen fixation (e.g., pcc/acc for carbon fixation, nifH for nitrogen fixation), carbon decomposition (e.g., amyA and pulA for labile carbon decomposition, mnp and lcc for recalcitrant carbon decomposition), and greenhouse gas emissions (e.g., mcrA for methane production, norB for nitrous oxide production) across five FACE sites. Also, the relative abundance of those key genes was generally increased and directionally associated with increased biomass, soil carbon decomposition, and soil moisture. In addition, a further literature survey of more disparate FACE experimental sites indicated increased biomass, soil carbon decay, nitrogen fixation, methane and nitrous oxide emissions, plant and soil carbon and nitrogen under eCO2. A conceptual framework was developed to link commonly responsive functional genes with ecosystem processes, such as pcc/acc vs. soil carbon storage, amyA/pulA/mnp/lcc vs. soil carbon decomposition, and nifH vs. nitrogen availability, suggesting that such common responses of microbial functional genes may have the potential to predict ecosystem functioning and sustainability in the future eCO2 environment.

In response to the existing problems that the federated learning might lead to the reduction of aggregation efficiency by handing the majority of irregular users and the leak of parameter privacy by adopting plaintext communication, a framework of privacy-preserving robust federated learning was proposed for ensuring the robustness of the irregular user based on the designed security division protocol. PPRFL could enable the model and its related information to aggregate in ciphertext on the edge server facilitate users to calculate the model reliability locally for reducing the additional communication overhead caused by the adoption of the security multiplication protocol in conventional methods, apart from lowering the high computational overhead resulted from homomorphic encryption with outsourcing computing to two edge servers. Based on this, user could calculate the loss value of the model through jointly using the verification sets issued by the edge server and that held locally after parameter updating of the local model. Then the model reliability could be dynamically updated as the model weight together with the historic information of the loss value. Further, the model weight was scaled under the guidance of prior knowledge, and the ciphertext model and ciphertext weight information are sent to the edge server to aggregate and update the global model parameters, ensuring that global model changes are contributed by high-quality data users, and improving the convergence speed. Through the security analysis of the Hybrid Argument model, the demonstration shows that PPRFL can effectively protect the privacy of model parameters and intermediate interaction parameters including user reliability. The experimental results show that the PPRFL scheme could still achieve the accuracy of 92% when all the participants in the federated aggregation task are irregular users, with the convergence efficiency 1.4 times higher than that of the PPFDL. Besides, the PPRFL scheme could still reach the accuracy of 89% when training data possessed by 80% of the users in the federated aggregation task were noise data, with the convergence efficiency 2.3 times higher than that of the PPFDL.

To tackle the problems of users not having intuitive cognition of the dynamic privacy changes contained in their sensing data in mobile edge crowdsensing (MECS) and lack of personalized privacy risk warning values in the data uploading stage, a dynamic privacy measurement (DPM) model was proposed. A structured representation of data obtained by a user participating in a sensing task was introduced and was transformed it into a numerical matrix. Then privacy attribute preference and timeliness were presented to quantify the dynamic privacy changes of data. With this, personalized privacy thresholds of users based on the numerical matrix were reasonably calculated. Finally, differential privacy processing was performed on the numerical matrix, and a model evaluation system was designed for the proposed model. The simulation results show that the DPM model was effective and practical. According to the given example, a data utility of approximately 0.7 can be achieved, and the degree of privacy protection can be significantly improved as the noise level increases, adapting to the MECS of IoT.

Enterocytozoon hepatopenaei (EHP) is a parasite in shrimp farming. EHP mainly parasitizes the hepatopancreas of shrimp, causing slow growth, which severely restricts the economic income of shrimp farmers. To explore the pathogenic mechanism of EHP, the host subcellular construction, molecular biological characteristics, and mitochondrial condition of Litopenaeus vannamei were identified using transmission electron microscopy (TEM), real-time qPCR, an enzyme assay, and flow cytometry. The results showed that EHP spores, approximately 1 μm in size, were located on the cytoplasm of the hepatopancreas. The number of mitochondria increased significantly, and mitochondria morphology showed a condensed state in the high-concentration EHP-infected shrimp by TEM observation. In addition, there were some changes in mitochondrial potential, but apoptosis was not significantly different in the infected shrimp. The qPCR results showed that the gene expression levels of hexokinase and pyruvate kinase related to energy metabolism were both upregulated in the diseased L. vannamei. Enzymatic activity showed hexokinase and lactate dehydrogenase were significantly increased in the shrimp infected with EHP, indicating EHP infection can increase the glycolysis process and decrease the oxidative phosphorylation process of L. vannamei. Previous transcriptomic data analysis results also support this conclusion. [ABSTRACT FROM AUTHOR]

Aiming at the information leakage problem in the process of deep neural network model calculation, a series of secure and efficient interactive computing protocols were designed between two non-collusive edge servers in combination with the additive secret sharing scheme. Since the nonlinear function cannot be split directly, a set of basic conversion protocols were proposed to realize the secure conversion of additive and multiplicative shares. After a few invokes, the power, comparison, exponential, logarithm, division and other low-level functions can be calculated securely. Due to the characteristics of data transfer and computation, the proposed protocols can be extended to array computation. Theoretical analysis ensures the correctness, efficiency and security of these protocols. The experimental results show that the error of these protocols is negligible, and the computational costs and communication overhead are better than the existing schemes.

In mobile crowd sensing(MCS), attackers can reconstruct the social circle among sensing users, who use the social association information among sensing users and the correlation between the sensing user’s identity and sensing data to further attack a social alliance. In order to tackle this issue, a fog-aided identity privacy protection scheme is proposed. Firstly, two fog nodes are introduced which located at the edge of the sensing terminal. The one is task allocation center (TC) for handling the reasonable allocation of sensing tasks, and the other is data center (DC) for calculating sensing data. Furthermore, differential privacy is employed for preventing attackers from acquiring the specific social association weight of sensing users. Finally, in order to prevent attackers from obtaining the sensing users' identity information and sensing data at the same time, sensing users use different blind identities to communicate with TC and DC. Security analysis indicates that the proposed scheme can ensure the security of identity privacy information of sensing users in the process of completing sensing tasks. Experimental results show that the proposed scheme can protect the social association information between sensing users, and has a low delay.