Search

Facial expressions can reflect people’s inner emotions to a certain extent, and studying facial expressions can help psychologists capture expression information in time and understand patients’ psychological changes quickly. In this paper, we establish a multi-channel convolutional neural network face expression recognition model based on the fusion of the attention mechanism. With the help of the attention mechanism and multi-channel convolutional neural network, we input expression images and perform average pooling and maximum pooling, output the features with high recognition after pooling, and identify the features with high recognition in expression images throughout the process. And with the help of multi-scale feature fusion, we improve the detection of subtle changes, such as the corners of the mouth and the eyes of the expression image target. The loss function is used to calculate the loss rate of facial expression images, which leads to the correct rate of facial expression recognition by a multi-channel convolutional neural network based on the fusion of attention mechanisms. It is demonstrated that the highest recognition correct rate of the multi-channel convolutional neural network faces expression recognition model with attention mechanism fusion is 93.56% on the FER2013 dataset, which is higher than that of the MHBP model by 23.2%. The highest correct recognition rate on the RAF-DB dataset is 91.34%, which is higher than the SR-VGG19 model by 19.39%. This shows that the multi-channel convolutional neural network face expression recognition based on the fusion of attention mechanisms improves the correct rate of facial expression recognition, which is beneficial to the research and development of psychology.

Abstract Hydrogen (H2) can protect against blood‒brain barrier (BBB) damage in sepsis-associated encephalopathy (SAE), but the mechanism is still unclear. We examined whether it is related to PPARα and its regulatory targets, ABC efflux transporters. After injection with DMSO/GW6471 (a PPARα inhibitor), the mice subjected to sham/caecal ligation and puncture (CLP) surgery were treated with H2 for 60 min postoperation. Additionally, bEnd.3 cells were grown in DMSO/GW6471-containing or saline medium with LPS. In addition to the survival rates, cognitive function was assessed using the Y-maze and fear conditioning tests. Brain tissues were stained with TUNEL and Nissl staining. Additionally, inflammatory mediators (TNF-α, IL-6, HMGB1, and IL-1β) were evaluated with ELISA, and PPARα, ZO-1, occludin, VE-cadherin, P-gp, BCRP and MRP2 were detected using Western blotting. BBB destruction was assessed by brain water content and Evans blue (EB) extravasation. Finally, we found that H2 improved survival rates and brain dysfunction and decreased inflammatory cytokines. Furthermore, H2 decreased water content in the brain and EB extravasation and increased ZO-1, occludin, VE-cadherin and ABC efflux transporters regulated by PPARα. Thus, we concluded that H2 decreases BBB permeability to protect against brain dysfunction in sepsis; this effect is mediated by PPARα and its regulation of ABC efflux transporters.

Security of computer information can be improved with the use of a network intrusion detection system. Since the network environment is becoming more complex, more and more new methods of attacking the network have emerged, making the original intrusion detection methods ineffective. Increased network activity also causes intrusion detection systems to identify errors more frequently. We suggest a new intrusion detection technique in this research that combines a Convolutional Neural Network (CNN) model with a Bi-directional Long Short-term Memory Network (BiLSTM) model for adding attention mechanisms. We distinguish our model from existing methods in three ways. First, we use the NCR-SMOTE algorithm to resample the dataset. Secondly, we use recursive feature elimination method based on extreme random tree to select features. Thirdly, we improve the profitability and accuracy of predictions by adding attention mechanism to CNN-BiLSTM. This experiment uses UNSW-UB15 dataset composed of real traffic, and the accuracy rate of multi-classification is 84.5 |$\%$|⁠ ; the accuracy rate of multi-classification in CSE-IC-IDS2018 dataset reached 98.3 |$\%$|⁠. [ABSTRACT FROM AUTHOR]

Spatial power load forecasting is crucial for power grid planning, generation planning, dispatching, efficient power utilization, and sustainable development. The integration of new energy sources and electric vehicles has significantly altered grid loads, increasing the complexity of spatial load forecasting. However, existing techniques fail to fully consider the temporal and spatial correlation characteristics of data, leading to challenges in data identification and summarization. This reduces load forecasting accuracy and prolongs prediction time. To address these issues, a spatial electric load forecasting method based on improved scale limited dynamic time warping (LDTW) and graph convolutional network (GCN) are proposed. Firstly, the improved scale LDTW is used to improve the clustering effect of K‐Mediods++, refine the type of load data, and make the subsequent model training more targeted. Secondly, the interconnections and distances of substations in a real network structure is used to build a graph model to capture the power load distribution. Finally, based on the clustering results and the graph model, GCN‐LSTM is used to construct the spatio‐temporal forecasting algorithm. The proposed algorithm is tested using load data from a region in Shanghai and compared with other advanced algorithms. Results show that the algorithm achieves higher prediction accuracy and efficiency. [ABSTRACT FROM AUTHOR]

Searchable asymmetric encryption (SAE) enables a client to search over a data owner's encrypted data. Nevertheless, state-of-the-art SAE schemes allow a data owner to specify access control policy for a client, while they have not considered the threat case of a malicious data owner. To address the problem, this work presents a non-interactive SAE scheme with bilateral access control: (i) allowing data owner and client to both specify policies toward the other party; (ii) allowing client to perform arbitrary boolean queries with sub-linear search complexity. Technically, we extend Cash et al.'s highly scalable SSE into an asymmetric setting and introduce the property of data owner authenticity. By refining identity-based matchmaking encryption, we formalize the syntax and security definition of our SAE with identity-based bilateral access control. Moreover, the security of the proposed SAE can be reduced to discrete logistic assumption and decisional bilinear Diffie–Hellman assumption. As an enhanced extension, we present a non-interactive multi-client SAE scheme with fuzzy identity-based bilateral access control. In addition, we implement the proposed schemes in real cloud platform and evaluate their performance on a real-world dataset. The result confirms that our SAE schemes achieve bilateral access control for both data owner and client with highly acceptable efficiency. [ABSTRACT FROM AUTHOR]