Explore the vast range of titles available.

Outlier detection is an important task in the field of data mining and a highly active area of research in machine learning. In industrial automation, datasets are often high-dimensional, meaning an effort to study all dimensions directly leads to data sparsity, thus causing outliers to be masked by noise effects in high-dimensional spaces. The “curse of dimensionality” phenomenon renders many conventional outlier detection methods ineffective. This paper proposes a new outlier detection algorithm called EOEH (Ensemble Outlier Detection Method Based on Information Entropy-Weighted Subspaces for High-Dimensional Data). First, random secondary subsampling is performed on the data, and detectors are run on various small-scale sub-samples to provide diverse detection results. Results are then aggregated to reduce the global variance and enhance the robustness of the algorithm. Subsequently, information entropy is utilized to construct a dimension-space weighting method that can discern the influential factors within different dimensional spaces. This method generates weighted subspaces and dimensions for data objects, reducing the impact of noise created by high-dimensional data and improving high-dimensional data detection performance. Finally, this study offers a design for a new high-precision local outlier factor (HPLOF) detector that amplifies the differentiation between normal and outlier data, thereby improving the detection performance of the algorithm. The feasibility of this algorithm is validated through experiments that used both simulated and UCI datasets. In comparison to popular outlier detection algorithms, our algorithm demonstrates a superior detection performance and runtime efficiency. Compared with the current popular, common algorithms, the EOEH algorithm improves the detection performance by 6% on average. In terms of running time for high-dimensional data, EOEH is 20% faster than the current popular algorithms.

The detection and defense of malicious attacks are critical to the proper functioning of network security. Due to the diversity and rapid updates of the attack methods used by attackers, traditional defense mechanisms have been challenged. In this context, a more effective method to predict vulnerabilities in network systems is considered an urgent need to protect network security. In this paper, we propose a formal modeling and analysis approach based on Petri net vulnerability exploitation. We used the Common Vulnerabilities and Exposures (CVE)-2021-3711 vulnerability source code to build a model. A patch model was built to address the problems of this model. Finally, the time injected by the actual attacker and the time simulated by the software were calculated separately. The results showed that the simulation time was shorter than the actual attack time, and ultra-real-time simulation could be achieved. By modeling the network system with this method, the model can be found to arrive at an illegitimate state according to the structure of Petri nets themselves and thus discover unknown vulnerabilities. This method provides a reference method for exploring unknown vulnerabilities.

The distribution of wireless network systems challenges the communication security of Internet of Things (IoT), and the IPv6 protocol is gradually becoming the main communication protocol under the IoT. The Neighbor Discovery Protocol (NDP), as the base protocol of IPv6, includes address resolution, DAD, route redirection and other functions. The NDP protocol faces many attacks, such as DDoS attacks, MITM attacks, etc. In this paper, we focus on the communication-addressing problem between nodes in the Internet of Things (IoT). We propose a Petri-Net-based NS flooding attack model for the flooding attack problem of address resolution protocols under the NDP protocol. Through a fine-grained analysis of the Petri Net model and attacking techniques, we propose another Petri-Net-based defense model under the SDN architecture, achieving security for communications. We further simulate the normal communication between nodes in the EVE-NG simulation environment. We implement a DDoS attack on the communication protocol by an attacker who obtains the attack data through the THC-IPv6 tool. In this paper, the SVM algorithm, random forest algorithm (RF) and Bayesian algorithm (NBC) are used to process the attack data. The NBC algorithm is proven to exhibit high accuracy in classifying and identifying data through experiments. Further, the abnormal data are discarded through the abnormal data processing rules issued by the controller in the SDN architecture, to ensure the security of communications between nodes.

In the field of music notation recognition, while the recognition technology for common notation systems such as staff notation has become quite mature, the recognition techniques for traditional Chinese notation systems like guqin tablature (jianzipu) and Kunqu opera gongchepu remain relatively underdeveloped. As an important carrier of China’s thousand-year musical culture, the digital preservation and inheritance of Kunqu opera’s Gongche notation hold significant cultural value and practical significance. By addressing the unique characteristics of Gongche notation, this study overcomes the limitations of Western staff notation recognition technologies. By constructing a deep learning model adapted to the morphology of Chinese character-style notation symbols, it provides technical support for establishing an intelligent processing system for Chinese musical documents, thereby promoting the innovative development and inheritance of traditional music in the era of artificial intelligence. This paper has constructed the LGRC2024 (Gong-che notation based on Lilu Qu Pu) dataset. It has also employed data augmentation operations such as image translation, rotation, and noise processing to enhance the diversity of the dataset. For the recognition of Gong-che notation, the YOLOv8 model was adopted, and the network performances of its lightweight (n) and medium-weight (m) versions were compared and analyzed. The superior-performing YOLOv8m was selected as the basic model. To further improve the model’s performance, SimAM, Triplet Attention, and Multi-scale Convolutional Attention Module (MSCAM) were introduced to optimize the model. The experimental results show that the accuracy of the basic YOLOv8m model increased from 65.9% to 78.2%. The improved models based on YOLOv8m achieved recognition accuracies of 80.4%, 81.8%, and 83.6%, respectively. Among them, the improved model with the MSCAM module demonstrated the best performance in all aspects.

A high-calorie diet and lack of exercise are the most important risk factors contributing to metabolic dysfunction-associated steatotic liver disease (MASLD) initiation and progression. The precise molecular mechanisms of mitochondrial function alteration during MASLD development remain to be fully elucidated. In this study, a total of 60 male C57BL/6J mice were maintained on a normal or amylin liver NASH (AMLN) diet for 6 or 10 weeks. Some of the mice were then subjected to voluntary wheel running, while the other mice were fed a normal or AMLN diet until 14 and 18 weeks. The results showed that hepatic lipid deposition and the PERK-eIF2α-ATF4 pathway were significantly increased with prolonged duration of AMLN diet. However, expression of mitochondrial unfolded protein response (UPRmt) genes and mitokine FGF21 secretion were significantly enhanced in the 14-week AMLN diet mice, but were markedly reduced with the excessive lipid deposition induced by longer AMLN diet. Additionally, the exercise intervention acts as a regulator to optimize UPRmt signal transduction and to enhance mitochondrial homeostasis by improving mitochondrial function, reversing the UPRmt activation pattern, and increasing FGF21 secretion, which plays a pivotal role in delaying the occurrence and development of MASLD.

Pavement roughness is a critical indicator for road maintenance decisions and driving safety assessment. Existing methods primarily rely on multi-source explicit features, which have limited capability in capturing implicit information such as spatial topology between road segments. Furthermore, their accuracy and stability remain insufficient in cross-regional and small-sample prediction scenarios. To address these limitations, we propose a Graph-Fused Natural Gradient Boosting framework (GF-NGB), which combines the spatial topology modeling capability of graph neural networks with the small-sample robustness of natural gradient boosting for high-precision cross-regional roughness prediction. The method first extracts an 18-dimensional set of multi-source features from the U.S. Long-Term Pavement Performance (LTPP) database and derives an 8-dimensional set of implicit spatial features using a graph neural network. These features are then concatenated and fed into a natural gradient boosting model, which is optimized by Optuna, to predict the dual objectives of left and right wheel-track roughness. To evaluate the generalization capability of the proposed method, we employ a spatially partitioned data split: the training set includes 1648 segments from Arizona, California, Florida, Ontario, and Missouri, while the test set comprises 330 segments from Manitoba and Nevada with distinct geographic and climatic conditions. Experimental results show that GF-NGB achieves the best performance on cross-regional tests, with average prediction accuracy improved by 1.7% and 3.6% compared to Natural Gradient Boosting (NGBoost) and a Graph Neural Network–Multilayer Perceptron hybrid model (GNN-MLP), respectively. This study reveals the synergistic effect of multi-source texture features and spatial topology information, providing a generalizable framework and technical pathway for cross-regional, small-sample intelligent pavement monitoring and smart maintenance.

Alzheimer's disease (AD) is one of the most prevalent forms of dementia globally and remains an incurable condition that often leads to death. PANoptosis represents an emerging paradigm in programmed cell death, integrating three critical processes: pyroptosis, apoptosis, and necroptosis. Studies have shown that apoptosis, necroptosis, and pyroptosis play important roles in AD development. Therefore, targeting PANoptosis genes might lead to novel therapeutic targets and clinically relevant therapeutic approaches. This study aims to identify different molecular subtypes of AD and potential drugs for treating AD based on PANoptosis. Differentially expressed PANoptosis genes associated with AD were identified via Gene Expression Omnibus (GEO) dataset GSE48350, GSE5281, and GSE122063. Least Absolute Shrinkage and Selection Operator (LASSO) regression was employed to construct a risk model linked to these PANoptosis genes. Consensus clustering analysis was conducted to define AD subtypes based on these genes. We further performed gene set variation analysis (GSVA), functional enrichment analysis, and immune cell infiltration analysis to investigate differences between the identified AD subtypes. Additionally, a protein-protein interaction (PPI) network was established to identify hub genes, and the DGIdb database was consulted to identify potential therapeutic compounds targeting these hub genes. Single-cell RNA sequencing analysis was utilized to assess differences in gene expression at the cellular level across subtypes. A total of 24 differentially expressed PANoptosis genes (APANRGs) were identified in AD, leading to the classification of two distinct AD subgroups. The results indicate that these subgroups exhibit varying disease progression states, with the early subtype primarily linked to dysfunctional synaptic signaling. Furthermore, we identified hub genes from the differentially expressed genes (DEGs) between the two clusters and predicted 38 candidate drugs and compounds for early AD treatment based on these hub genes. Single-cell RNA sequencing analysis revealed that key genes associated with the early subtype are predominantly expressed in neuronal cells, while the differential genes for the metabolic subtype are primarily found in endothelial cells and astrocytes. In summary, we identified two subtypes, including the AD early synaptic abnormality subtype as well as the immune-metabolic subtype. Additionally, ten hub genes, SLC17A7, SNAP25, GAD1, SLC17A6, SLC32A1, PVALB, SYP, GRIN2A, SLC12A5, and SYN2, were identified as marker genes for the early subtype. These findings may provide valuable insights for the early diagnosis of AD and contribute to the development of innovative therapeutic strategies.

Recent studies demonstrate that algorithmic music attracted global attention not only because of its amusement but also its considerable potential in the industry. Thus, the yield increased academic numbers spinning around on topics of algorithm music generation. The balance between mathematical logic and aesthetic value is important in music generation. To maintain this balance, we propose a research method based on a three-dimensional temporal convolutional attention neural network. This method uses a self-collected traditional Chinese pentatonic symbolic music dataset. It combines clustering algorithms and deep learning-related algorithms to construct a three-dimensional sequential convolutional generation model 3D-SCN, a three-dimensional temporal convolutional attention model BoYaTCN. We trained both of them to generate traditional Chinese pentatonic scale music that considers both overall temporal creativity and local musical semantics. Then, we conducted quantitative and qualitative evaluations of the generated music. The experiment demonstrates that BoYaTCN achieves the best results, with a prediction accuracy of 99.12%, followed by 3D-SCN with a prediction accuracy of 99.04%. We have proven that the proposed model can generate folk music with a beautiful melody, harmonious coherence, and distinctive traditional Chinese pentatonic features, and it also conforms to certain musical grammatical characteristics.

Obtaining the liquid storage state of oil wells in real time is very important for oilfield production. In this paper, under the premise of fully considering the transformation factors of full-pumping and nonfull-pumping states of oil wells, submergence depth models suitable for full- and nonfull-pumping wells are constructed. To reduce the application complexity of the models, parameter-reduction processing is performed to enhance the usability of the models. By analyzing the change trend of the submergence depth during the rising, maintaining, and falling of the oil well in the full-pumping state and nonfull-pumping state models, the judgment criteria for the transition of the oil well state are provided. On this basis, the application methods of nonlinear interpolation and least squares curve-fitting numerical solutions of submergence depth models are studied, and the unique existence of the solution of the corresponding one-variable nonlinear characteristic equation in the (0, 1) open interval is proven. Finally, the error estimation of the numerical solution is carried out, the calculation formula of the number of iterations for the numerical solution of the dichotomy is provided, and the error of the relevant numerical solution is verified.

With the development of society, the intangible cultural heritage of Chinese Nüshu is in danger of extinction. To promote the research and popularization of traditional Chinese culture, we use deep learning to automatically detect and recognize handwritten Nüshu characters. To address difficulties such as the creation of a Nüshu character dataset, uneven samples, and difficulties in character recognition, we first build a large-scale handwritten Nüshu character dataset, HWNS2023, by using various data augmentation methods. This dataset contains 5500 Nüshu images and 1364 labeled character samples. Second, in this paper, we propose a two-stage scheme model combining GoogLeNet-tiny and YOLOv5-CBAM (SGooTY) for Nüshu recognition. In the first stage, five basic deep learning models including AlexNet, VGGNet16, GoogLeNet, MobileNetV3, and ResNet are trained and tested on the dataset, and the model structure is improved to enhance the accuracy of recognising handwritten Nüshu characters. In the second stage, we combine an object detection model to re-recognize misidentified handwritten Nüshu characters to ensure the accuracy of the overall system. Experimental results show that in the first stage, the improved model achieves the highest accuracy of 99.3% in recognising Nüshu characters, which significantly improves the recognition rate of handwritten Nüshu characters. After integrating the object recognition model, the overall recognition accuracy of the model reached 99.9%.