Narrow Results

It is important to understand the behavior of an information network and its features. In this research, we explore this idea by applying a multidimensional data analysis system, a system that is significant in enhancing data intrusion detection. To accomplish this, we gather data from various sources, like network traffic, user behavior, and system logs. Using the examples mentioned, the proposed system detects and prevents cyber threats more accurately. Artificial Intelligence techniques including deep learning, clustering, and principal component analysis (PCA), are used. These are essential in analyzing complex patterns within the data and enabling the detection of sophisticated and evolving intrusion techniques. Multidimensional data allows for the capture of intricate, non-linear relationships. This improves the system’s ability to differentiate between normal and abnormal activities. Real-time data processing and AI-driven algorithms enhance detection speed and enable faster responses to potential intrusions. We tested this system on benchmark datasets. Results showed significant improvements in detection rates and a reduction in false positives compared to traditional NIDS. The integration of AI gave us a more adaptive and scalable approach to intrusion detection. Also, it allowed the system to learn from new attack patterns and continuously refine its capabilities.

With the rapid expansion of power grids and increasing user demand, effectively identifying and monitoring abnormal electricity consumption have become crucial for ensuring grid stability and operational efficiency. Traditional anomaly detection methods often struggle with scalability and accuracy, particularly as the volume and density of electricity data grow. To address these challenges, this paper introduces a novel electricity anomaly monitoring framework that integrates real-time data acquisition and advanced classification modeling techniques. Our approach leverages a parallel classification algorithm designed to efficiently handle large datasets and detect anomalies with high accuracy. Key features of abnormal user data are extracted using information entropy, and electricity consumption data are continuously collected through a wireless network. The proposed method then preprocesses and classifies the data, applying a random forest model to detect anomalies and monitor usage patterns. Experimental results indicate that our approach significantly enhances both the accuracy and efficiency of electricity anomaly detection, demonstrating its robustness and potential for large-scale deployment in power grid systems.

Efficient power networks are crucial for modern society, ensuring reliable electricity supply to homes, businesses, and industries while supporting economic growth and technological advancements. They also contribute to environmental sustainability by facilitating the transition to renewable energy sources and reducing environmental impact. Moreover, efficient power networks play a critical role in disaster response, healthcare delivery, and communication systems. Given the complexities inherent in power networks, effective anomaly detection is essential to prevent disruptions like power outages and economic losses. To address this, a novel artificial intelligence-based anomaly detection and localization approach is proposed. This approach involves three key steps: preprocessing, feature extraction, and anomaly detection and localization. In the preprocessing step, the signal decomposition technique breaks down the input signal into smaller segments. Statistical features are then extracted from these segments in the feature extraction step. These features are inputted into an Improved Deep Neural Network (IDNN) model for anomaly detection. Finally, an Enhanced Coati Optimization Algorithm (ECOA)-based Support Vector Regression (SVR) model is utilized to locate the exact location of the detected anomaly. The proposed approach is evaluated through various analyses to demonstrate its effectiveness in enhancing situational awareness and resilience, thereby ensuring the reliability and stability of complex power networks.

In order to improve the service quality of the power grid and ensure the safe and stable operation of the power grid, an in-depth study is carried out on the automatic detection and positioning of abnormal power metering. First, the data clustering algorithm is used to cluster the electric energy metering data. Next, the graph convolutional neural network algorithm is used to automatically detect and locate the abnormal data of electric energy metering. A graph is constructed, and the subgraphs are divided. These subgraphs are then sent to the graph convolutional neural network for processing. Feature extraction is carried out on the points and edges of the graph. This process completes the automatic detection and location of the electric energy metering anomaly based on the graph convolutional neural network. The experimental results show that the proposed algorithm can effectively improve the efficiency and accuracy of automatic detection and positioning of electric energy metering anomalies, and can better meet the needs of actual power work.

The incorporation of Information and Communication Technologies (ICT) into traditional power grids has transformed them into smart grids, revolutionizing energy management systems. At the core of this transformation are Intelligent Electronic Devices (IEDs), which provide essential data for key Energy Management System (EMS) applications, such as state estimation and optimal power flow. IEDs are critical for ensuring the stability and security of smart grid operations, but they are also vulnerable to various anomalies, including infrastructure faults, equipment malfunctions, energy theft, and cyberattacks. Detecting these anomalies is vital to maintaining the reliability of smart grid systems and preventing potential threats to national security. This study introduces a statistical data-driven framework designed to detect and explain anomalies in IED-based smart grid systems. The framework includes a preprocessing module for ensuring high-quality input data and an anomaly detection module that prioritizes interpretability and explainability. Using methods like the Gaussian Mixture Model (GMM), Kalman Filter (KF), and ExtraTree Classifier, the framework achieves 99% accuracy in anomaly detection and 88% accuracy in classifying events as either natural occurrences or cyberattacks.

With the rapid development of the Internet, network user behavior data shows explosive growth. How to accurately identify abnormal users from massive data is of great significance for maintaining network security and preventing network crimes. This study first comprehensively collected multimodal data from users, including text data, behavioral data and possible image data. These data form the foundation of the research and provide rich materials for subsequent feature extraction and model construction. In the feature extraction stage, we adopted different processing methods for data of different modalities. For text data, natural language processing techniques are used to extract features such as keywords and emotional tendencies. Mining patterns and anomalies in user behavior through statistical analysis, time series analysis and other methods for behavioral data. By using computer vision technology to extract image features from image data, these features collectively constitute a multimodal feature set of user behavior. The experimental results show that the anomaly network user detection method based on multimodal data fusion and hybrid neural networks (HNN) has high accuracy and robustness. Compared with single modal data or traditional detection methods, this method shows significant advantages in identifying abnormal users. In addition, this method can provide rich user behavior characteristic information, which provides strong support for network security analysis.

Predictive maintenance (PdM) helps organizations to reduce equipment downtime, optimize maintenance schedules, and enhance operational efficiency. By leveraging machine learning algorithms to predict when equipment failure will likely occur, maintenance teams can proactively schedule maintenance activities and prevent unexpected breakdowns. Fault detection and diagnosis are essential components of PdM. Fault detection involves analyzing sensor data collected from equipment to identify deviations from normal behavior. Diagnosis, however, involves identifying the root cause of a fault or failure. A dataset of an industrial asset is used to evaluate the proposed study. K-means clustering anomaly detection approach is employed. Implementing machine learning (ML)-based fault categorization approaches revealed that Random Forest had the best results. Significant progress has been made in fault detection and diagnosis using ML, but the degree of their explainability is significantly limited by the “black-box” character of some ML techniques. Less emphasis has been placed on explainable artificial intelligence (XAI) approaches in maintenance. Therefore, the XAI tools, SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) have been used to acknowledge the extent of the variables to analyze the influence of respective features. A stability metric has been included to improve the explanation’s overall quality. The findings of this paper suggest that the utilization of XAI can offer significant contributions in terms of insights and solutions for addressing critical maintenance issues.

The stability test of multi-station precision machining fixtures can improve the safety of multi-station precision machining fixture. This paper puts forward a stability test method of multi-station precision machining fixtures based on the pressure pulsation amplitude test. The fundamental frequency characteristic detection of airflow pulsation is adopted to collect the stability characteristic parameters of multi-station precision machining machine clamps, and the wave number analysis model of airflow pulsation in the stable state of multi-station precision machining machine clamps is constructed. The time domain characteristic parameter analysis and spectrum gain control are adopted to realize the stability condition parameter analysis and spectrum characteristic decomposition of multi-station precision machining machine clamps. The end-face pulsation attenuation characteristic method of detection is followed to realize noise reduction of the stable air pulsation attenuation signal of multi-station precision machining fixture, and the rotor meshing frequency component of the stable air pulsation attenuation signal of multi-station precision machining fixture is extracted. The multi-scale wavelet air pulsation attenuation signal decomposition is adopted to realize the correlation spectrum characteristic detection of the signal, and the multi-dimensional space fusion is employed to realize signal characteristic clustering, thus realizing the stability detection of multi-station precision machining fixture. The test output shows that the proposed method has better convergence and strong feature clustering in the stability detection of multi-station precision machining machine fixtures, which improves the confidence level of multi-station precision machining machine fixture failure instability detection.

The significance of shielding university network data has improved with the rapid digitalization of the educational process. Distributed security and elevated perimeter defense threats, which place the bigger data center in danger, are the challenges of edge security. This research focuses on creating edge computing (EC) based anomaly detection and security protection that is specially designed for university networks. In this study, we suggested a novel social spider optimization enhanced long-short-term memory (SSO-ELSTM) for attack detection of university network data and security. In the suggested approach for attack detection, we have employed the CTU dataset. The data were preprocessed using Z-score normalization. Then the preprocessed data were used to extract the features for principle component analysis (PCA) using dimensional reduction. The proposed method is implemented using Python software. We compare the suggested method with other existing methods. The outcomes demonstrate that the suggested approach outperforms the other method in terms of accuracy, precision, recall, F1 score and error rate. This study presents a framework for using edge computing and sophisticated anomaly detection methods to enhance university network security; the outcome highlights practicality and efficiency of edge-based solutions for shielding sensitive university network information.

Power system security is a critical concern for modern energy grids, where both natural disturbances and human-made events can lead to system failures or significant disruptions. Effective monitoring is essential to maintain stability, reduce risks, and ensure the system’s reliable operation. This study aims to develop an advanced security monitoring framework for power systems using quantum computing to enhance anomaly detection and establish an early warning system. It focuses on identifying and mitigating disturbances through a robust model that improves system security and resilience. The dataset used for this study is sourced from phasor measurement units (PMUs), which are critical for identifying anomalies. The z-score normalization approach was used to preprocess the dataset. Wavelet transforms and statistical techniques are combined to extract key characteristics, which aids in finding patterns and trends that were necessary for spotting abnormalities. The stochastic fractal search-driven quantum-enhanced scalable support vector regression (SFS-QE-SSVR) is designed to improve prediction accuracy by optimizing support vector regression models. The quantum enhancement boosts the algorithm’s performance by accelerating the search process for optimal parameters. The system detects anomalies based on deviations from normal patterns and triggers early warnings, allowing operators to take preventive actions before critical failures occur. The model demonstrates superior accuracy in detecting anomalies and cyber-attacks, outperforming traditional machine-learning methods. In a comparative analysis, the suggested method is assessed with various evaluation measures, such as $F1$-score (98.15%), recall (97.10%), precision (98.35%) and accuracy (98%). The result demonstrated that the SFS-QE-SSVR method improves the prediction accuracy by optimizing support vector regression models. The proposed quantum-enhanced approach significantly enhances power system security and efficiency, providing a reliable solution for modern energy grids.

Diagnosis of learning difficulties is a challenging goal. There are huge number of factors involved in the evaluation procedure that present high variance among the population with the same difficulty. Diagnosis is usually performed by scoring subjects according to results obtained in different neuropsychological (performance-based) tests specifically designed to this end. One of the most frequent disorders is developmental dyslexia (DD), a specific difficulty in the acquisition of reading skills not related to mental age or inadequate schooling. Its prevalence is estimated between 5% and 12% of the population. Traditional tests for DD diagnosis aim to measure different behavioral variables involved in the reading process. In this paper, we propose a diagnostic method not based on behavioral variables but on involuntary neurophysiological responses to different auditory stimuli. The experiments performed use electroencephalography (EEG) signals to analyze the temporal behavior and the spectral content of the signal acquired from each electrode to extract relevant (temporal and spectral) features. Moreover, the relationship of the features extracted among electrodes allows to infer a connectivity-like model showing brain areas that process auditory stimuli in a synchronized way. Then an anomaly detection system based on the reconstruction residuals of an autoencoder using these features has been proposed. Hence, classification is performed by the proposed system based on the differences in the resulting connectivity models that have demonstrated to be a useful tool for differential diagnosis of DD as well as a method to step towards gaining a better knowledge of the brain processes involved in DD. The results corroborate that nonspeech stimulus modulated at specific frequencies related to the sampling processes developed in the brain to capture rhymes, syllables and phonemes produces effects in specific frequency bands that differentiate between controls and DD subjects. The proposed method showed relatively high sensitivity above 0.6, and up to 0.9 in some of the experiments.

Image anomaly detection is an application-driven problem where the aim is to identify novel samples, which differ significantly from the normal ones. We here propose Pyramidal Image Anomaly DEtector (PIADE), a deep reconstruction-based pyramidal approach, in which image features are extracted at different scale levels to better catch the peculiarities that could help to discriminate between normal and anomalous data. The features are dynamically routed to a reconstruction layer and anomalies can be identified by comparing the input image with its reconstruction. Unlike similar approaches, the comparison is done by using structural similarity and perceptual loss rather than trivial pixel-by-pixel comparison. The proposed method performed at par or better than the state-of-the-art methods when tested on publicly available datasets such as CIFAR10, COIL-100 and MVTec.

Image anomaly detection consists in detecting images or image portions that are visually different from the majority of the samples in a dataset. The task is of practical importance for various real-life applications like biomedical image analysis, visual inspection in industrial production, banking, traffic management, etc. Most of the current deep learning approaches rely on image reconstruction: the input image is projected in some latent space and then reconstructed, assuming that the network (mostly trained on normal data) will not be able to reconstruct the anomalous portions. However, this assumption does not always hold. We thus propose a new model based on the Vision Transformer architecture with patch masking: the input image is split in several patches, and each patch is reconstructed only from the surrounding data, thus ignoring the potentially anomalous information contained in the patch itself. We then show that multi-resolution patches and their collective embeddings provide a large improvement in the model’s performance compared to the exclusive use of the traditional square patches. The proposed model has been tested on popular anomaly detection datasets such as MVTec and head CT and achieved good results when compared to other state-of-the-art approaches.

Considering the 2030 United Nations intent of world connection, Cyber Intelligence becomes the main area of the human dimension able of inflicting changes in geopolitical dynamics. In cyberspace, the new battlefield is the mind of people including new weapons like abuse of social media with information manipulation, deception by activists and misinformation. In this paper, a Sentiment Analysis system with Anomaly Detection (SAAD) capability is proposed. The system, scalable and modular, uses an OSINT-Deep Learning approach to investigate on social media sentiment in order to predict suspicious anomaly trend in Twitter posts. Anomaly detection is investigated with a new semi-supervised process that is able to detect potentially dangerous situations in critical areas. The main contributions of the paper are the system suitability for working in different areas and domains, the anomaly detection procedure in sentiment context and a time-dependent confusion matrix to address model evaluation with unbalanced dataset. Real experiments and tests were performed on Sahel Region. The detected anomalies in negative sentiment have been checked by experts of Sahel area, proving true links between the models results and real situations observable from the tweets.

While the brain connectivity network can inform the understanding and diagnosis of developmental dyslexia, its cause–effect relationships have not yet enough been examined. Employing electroencephalography signals and band-limited white noise stimulus at 4.8 Hz (prosodic-syllabic frequency), we measure the phase Granger causalities among channels to identify differences between dyslexic learners and controls, thereby proposing a method to calculate directional connectivity. As causal relationships run in both directions, we explore three scenarios, namely channels’ activity as sources, as sinks, and in total. Our proposed method can be used for both classification and exploratory analysis. In all scenarios, we find confirmation of the established right-lateralized Theta sampling network anomaly, in line with the assumption of the temporal sampling framework of oscillatory differences in the Theta and Gamma bands. Further, we show that this anomaly primarily occurs in the causal relationships of channels acting as sinks, where it is significantly more pronounced than when only total activity is observed. In the sink scenario, our classifier obtains 0.84 and 0.88 accuracy and 0.87 and 0.93 AUC for the Theta and Gamma bands, respectively.

In this paper, we develop the so-called variable projection support vector machine (VP-SVM) algorithm that is a generalization of the classical SVM. In fact, the VP block serves as an automatic feature extractor to the SVM, which are trained simultaneously. We consider the primal form of the arising optimization task and investigate the use of nonlinear kernels. We show that by choosing the so-called adaptive Hermite function system as the basis of the orthogonal projections in our classification scheme, several real-world signal processing problems can be successfully solved. In particular, we test the effectiveness of our method in two case studies corresponding to anomaly detection. First, we consider the detection of abnormal peaks in accelerometer data caused by sensor malfunction. Then, we show that the proposed classification algorithm can be used to detect abnormalities in ECG data. Our experiments show that the proposed method produces comparable results to the state-of-the-art while retaining desired properties of SVM classification such as light weight architecture and interpretability. We implement the proposed method on a microcontroller and demonstrate its ability to be used for real-time applications. To further minimize computational cost, discrete orthogonal adaptive Hermite functions are introduced for the first time.

While many seizure detection methods have demonstrated great accuracy, their training necessitates a substantial volume of labeled data. To address this issue, we propose a novel method for unsupervised seizure anomaly detection called SAnoDDPM, which uses denoising diffusion probabilistic models (DDPM). We designed a novel pipeline that uses a variable lower bound on Markov chains to identify potential values that are unlikely to occur in anomalous data. The model is first trained on normal data, then anomalous data is input to the trained model. The model resamples the anomalous data and converts it to normal data. Finally, the presence of seizures can be determined by comparing the before and after data. Moreover, the input 2D spectrograms are encoded into vector-quantized representations, which enables powerful and efficient DDPM while maintaining its quality. Experimental comparisons on the publicly available datasets, CHB-MIT and TUH, show that our method delivers better results, significantly reduces inference time, and is suitable for deployment in a clinical environments. As far as we are aware, this is the first DDPM-based method for seizure anomaly detection. This novel approach significantly contributes to the progression of seizure detection algorithms, thereby augmenting their practicality in clinical settings.

This research presents a robust adversarial method for anomaly detection in real-world scenarios, leveraging the power of generative adversarial neural networks (GANs) through cycle consistency in reconstruction error. Traditional approaches often falter due to high variance in class-wise accuracy, rendering them ineffective across different anomaly types. Our proposed model addresses these challenges by introducing an innovative flow of information in the training procedure and integrating it as a new discriminator into the framework, thereby optimizing the training dynamics. Furthermore, it employs a supplementary distribution in the input space to steer reconstructions toward the normal data distribution. This adjustment distinctly isolates anomalous instances and enhances detection precision. Also, two unique anomaly scoring mechanisms were developed to augment detection capabilities. Comprehensive evaluations on six varied datasets have confirmed that our model outperforms one-class anomaly detection benchmarks. The implementation is openly accessible to the academic community, available on Github.^a

In this paper we assess the effect that clustering pixels into spectrally-similar background types, for example, soil, vegetation, and water in hyperspectral visible/near-IR/SWIR imagery, prior to applying a detection methodology has on material detection statistics. Specifically, we examine the effects of data segmentation on two statistically-based detection metrics, the Subspace Generalized Likelihood Ratio Test (Subspace GLRT) and the Adaptive Cosine Estimator (ACE), applied to a publicly-available AVIRIS datacube augmented with a synthetic material spectrum in selected pixels. The use of synthetic spectrum-augmented data enables quantitative comparison of Subspace-GLRT and ACE using Receiver Operating Characteristic (ROC) curves. For all cases investigated, Receiver Operating Characteristic (ROC) curves generated using ACE were as good as or superior to those generated using Subspace-GLRT. The favorability of ACE over Subspace-GLRT was more pronounced as the synthetic spectrum mixing fraction decreased. For probabilities of detection in the range of 50-80%, segmentation reduced the probability of false alarm by a factor of 3–5 when using ACE. In contrast, segmentation had no apparent effect on detection statistics using Subspace-GLRT, in this example.

The stand-off detection classification by laser induced fluorescence is the objective of the Biosense project. The sensor will perform the monitoring of a defined area around its location using an elastic lidar detector for particles cloud. The detection of cloud will trigger fluorescence probing of the cloud. To perform this task the area fluorescence background will be monitored in order to evaluate if a return signal changed. Using a simple signal model built with experimental data, we designed a detection and monitoring procedure for the fluorescence at a single location. Signal simulations have been performed to verify the operation of the system. The results of the simulation indicate the system is able to detect anomaly with small contrast between a signal and the background. The results will have to be extended to area surveillance and a more complete signal model for various environments in natural conditions is required