Application of Feedforward Neural Network and SPT Results in the Estimation of Seismic Soil Liquefaction Triggering

Soil liquefaction is a dangerous phenomenon for structures that lose their shear strength and soil resistance, occurring during seismic shocks such as earthquakes or sudden stress conditions. Determining the liquefaction and nonliquefaction capacity of soil is a difficult but necessary job when constructing structures in earthquake zones. Usually, the possibility of soil liquefaction is determined by laboratory tests on soil samples subjected to dynamic loads, and this is time-consuming and costly. Therefore, this study focuses on the development of a machine learning model called a Forward Neural Network (FNN) to estimate the activation of soil liquefaction under seismic condition. The database is collected from the published literature, including 270 liquefaction cases and 216 nonliquefaction case histories under different geological conditions and earthquakes used for construction and confirming the model. The model is built and optimized for hyperparameters based on a technique known as random search (RS). Then, the L2 regularization technique is used to solve the overfitting problem of the model. The analysis results are compared with a series of empirical formulas as well as some popular machine learning (ML) models. The results show that the RS-L2-FNN model successfully predicts soil liquefaction with an accuracy of 90.33% on the entire dataset and an average accuracy of 88.4% after 300 simulations which takes into account the random split of the datasets. Compared with the empirical formulas as well as other machine learning models, the RS-L2-FNN model shows superior performance and solves the overfitting problem of the model. In addition, the global sensitivity analysis technique is used to detect the most important input characteristics affecting the activation prediction of liquefied soils. The results show that the corrected SPT resistance (N1)60 is the most important input variable, affecting the determination of the liquefaction capacity of the soil. This study provides a powerful tool that allows rapid and accurate prediction of liquefaction based on several basic soil properties.

Can Recurrent Neural Networks Predict Inflation in Euro Zone as Good as Professional Forecasters?

This paper investigates whether a specific type of a recurrent neural network, in particular Jordan neural network (JNN), captures the expected inflation better than commonly used feedforward neural networks and traditional parametric time-series models. It also considers competing survey-based and model-based expected inflation towards ex-post actual inflation to find whose predictions are more accurate; predictions from survey respondents or forecasting modelers. Further, it proposes neural network modelling strategy when dealing with nonstationary time-series which exhibit long-memory property and nonlinear dependence with respect to lagged inputs and exogenous inputs as well. Following this strategy, overfitting problem was reduced until no improvement in forecasting accuracy of expected inflation is achieved. The main finding is that JNN predicts inflation in euro zone quite accurately within forecasting horizon of 2 years. Regarding rational expectation principle we have found a set of demand-pull and cost-push inflation characteristics as exogenous inputs which helps in reducing overfitting problem of recurrent neural network even more. The sample includes euro zone aggregated monthly observations from January 2000 to December 2019. The results also confirm that inflation expectations obtained from JNN are consistent with Survey of professional forecasters (SPF), and thus, monetary policy makers can use JNN as a complementary tool in shortcomings of other inflation expectations measures.

Network security situation prediction based on feature separation and dual attention mechanism

With the development of smart cities, network security has become more and more important. In order to improve the safety of smart cities, a situation prediction method based on feature separation and dual attention mechanism is presented in this paper. Firstly, according to the fact that the intrusion activity is a time series event, recurrent neural network (RNN) or RNN variant is used to stack the model. Then, we propose a feature separation method, which can alleviate the overfitting problem and reduce cost of model training by keeping the dimension unchanged. Finally, limited attention is proposed according to global attention. We sum the outputs of the two attention modules to form a dual attention mechanism, which can improve feature representation. Experiments have proved that compared with other existing prediction algorithms, the method has higher accuracy in network security situation prediction. In other words, the technology can help smart cities predict network attacks more accurately.

High Performance DeepFake Video Detection on CNN-Based with Attention Target-Specific Regions and Manual Distillation Extraction

The rapid development of deep learning models that can produce and synthesize hyper-realistic videos are known as DeepFakes. Moreover, the growth of forgery data has prompted concerns about malevolent intent usage. Detecting forgery videos are a crucial subject in the field of digital media. Nowadays, most models are based on deep learning neural networks and vision transformer, SOTA model with EfficientNetB7 backbone. However, due to the usage of excessively large backbones, these models have the intrinsic drawback of being too heavy. In our research, a high performance DeepFake detection model for manipulated video is proposed, ensuring accuracy of the model while keeping an appropriate weight. We inherited content from previous research projects related to distillation methodology but our proposal approached in a different way with manual distillation extraction, target-specific regions extraction, data augmentation, frame and multi-region ensemble, along with suggesting a CNN-based model as well as flexible classification with a dynamic threshold. Our proposal can reduce the overfitting problem, a common and particularly important problem affecting the quality of many models. So as to analyze the quality of our model, we performed tests on two datasets. DeepFake Detection Dataset (DFDC) with our model obtains 0.958 of AUC and 0.9243 of F1-score, compared with the SOTA model which obtains 0.972 of AUC and 0.906 of F1-score, and the smaller dataset Celeb-DF v2 with 0.978 of AUC and 0.9628 of F1-score.

Feature Merged Network for Oil Spill Detection using SAR Images

The frequency of marine oil spills has increased in recent years. The growing exploitation of marine oil and continuous increase in marine crude oil transportation has caused tremendous damage to the marine ecological environment. Using synthetic aperture radar (SAR) images to monitor marine oil spills can help control the spread of oil spill pollution over time and reduce the economic losses and environmental pollution caused by such spills. However, it is a significant challenge to distinguish between oil-spilled areas and oil-spill-like in SAR images. Semantic segmentation models based on deep learning have been used in this field to address this issue. In addition, this study is dedicated to improving the accuracy of the U-Shape Network (UNet) model in identifying oil spill areas and oil-spill-like areas and alleviating the overfitting problem of the model; a feature merge network (FMNet) is proposed for image segmentation. The global features of SAR image, which are high-frequency component in the frequency domain and represents the boundary between categories, are obtained by a threshold segmentation method. This can weaken the impact of spot noise in SAR image. Then high-dimensional features are extracted from the threshold segmentation results using convolution operation. These features are superimposed with to the down sampling and combined with the high-dimensional features of original image. The proposed model obtains more features, which allows the model to make more accurate decisions. The overall accuracy of the proposed method increased by 1.82% and reached 61.90% compared with the UNet. The recognition accuracy of oil spill areas and oil-spill-like areas increased by approximately 3% and reached 56.33%. The method proposed in this paper not only improves the recognition accuracy of the original model, but also alleviates the overfitting problem of the original model and provides a more effective monitoring method for marine oil spill monitoring. More importantly, the proposed method provides a design principle that opens up new development ideas for the optimization of other deep learning network models.

Addressing Overfitting Problem in Deep Learning-Based Solutions for Next Generation Data-Driven Networks

Next-generation networks are data-driven by design but face uncertainty due to various changing user group patterns and the hybrid nature of infrastructures running these systems. Meanwhile, the amount of data gathered in the computer system is increasing. How to classify and process the massive data to reduce the amount of data transmission in the network is a very worthy problem. Recent research uses deep learning to propose solutions for these and related issues. However, deep learning faces problems like overfitting that may undermine the effectiveness of its applications in solving different network problems. This paper considers the overfitting problem of convolutional neural network (CNN) models in practical applications. An algorithm for maximum pooling dropout and weight attenuation is proposed to avoid overfitting. First, design the maximum value pooling dropout in the pooling layer of the model to sparse the neurons and then introduce the regularization based on weight attenuation to reduce the complexity of the model when the gradient of the loss function is calculated by backpropagation. Theoretical analysis and experiments show that the proposed method can effectively avoid overfitting and can reduce the error rate of data set classification by more than 10% on average than other methods. The proposed method can improve the quality of different deep learning-based solutions designed for data management and processing in next-generation networks.

Skip-YOLO: Domestic Garbage Detection Using Deep Learning Method in Complex Multi-scenes

In In the field of garbage intelligent identification, similar garbage are difficult to be effectively detected due to different kinds of characteristics. This paper proposes a Skip-YOLO model for garbage detection in real life through the visual analysis of feature mapping in different neural networks. First of all, the receptive field of the model is enlarged through the large-size convolution kernel, which enhanced the shallow information of images. Secondly, the high-dimensional feature mappings of garbage is extracted by dense convolutional blocks. The sensitivity of similar features in the same type of garbage is enhanced by strengthening the sharing of shallow low semantics and deep high semantics information. Finally, the multi-scale high-dimensional feature mappings is integrated and sent to the YOLO layer to predict the type and location of garbage. Experimental results show that compared with the YOLOv3, the overall detection precision is increased by 22.5%, and the average recall rate is increased by 18.6%. In qualitative comparison, it successfully detects domestic garbage in complex multi-scenes. In addition, our approach alleviates the overfitting problem of deep residual blocks. The application case of waste sorting production line is used to further highlight the model generalization performance of our method.

An Improved Vulnerability Exploitation Prediction Model with Novel Cost Function and Custom Trained Word Vector Embedding

Successful cyber-attacks are caused by the exploitation of some vulnerabilities in the software and/or hardware that exist in systems deployed in premises or the cloud. Although hundreds of vulnerabilities are discovered every year, only a small fraction of them actually become exploited, thereby there exists a severe class imbalance between the number of exploited and non-exploited vulnerabilities. The open source national vulnerability database, the largest repository to index and maintain all known vulnerabilities, assigns a unique identifier to each vulnerability. Each registered vulnerability also gets a severity score based on the impact it might inflict upon if compromised. Recent research works showed that the cvss score is not the only factor to select a vulnerability for exploitation, and other attributes in the national vulnerability database can be effectively utilized as predictive feature to predict the most exploitable vulnerabilities. Since cybersecurity management is highly resource savvy, organizations such as cloud systems will benefit when the most likely exploitable vulnerabilities that exist in their system software or hardware can be predicted with as much accuracy and reliability as possible, to best utilize the available resources to fix those first. Various existing research works have developed vulnerability exploitation prediction models by addressing the existing class imbalance based on algorithmic and artificial data resampling techniques but still suffer greatly from the overfitting problem to the major class rendering them practically unreliable. In this research, we have designed a novel cost function feature to address the existing class imbalance. We also have utilized the available large text corpus in the extracted dataset to develop a custom-trained word vector that can better capture the context of the local text data for utilization as an embedded layer in neural networks. Our developed vulnerability exploitation prediction models powered by a novel cost function and custom-trained word vector have achieved very high overall performance metrics for accuracy, precision, recall, F1-Score and AUC score with values of 0.92, 0.89, 0.98, 0.94 and 0.97, respectively, thereby outperforming any existing models while successfully overcoming the existing overfitting problem for class imbalance.