Search

Submitted by Isabel Nazareth (iqn@lisboa.ucp.pt) on 2021-11-03T15:57:52Z No. of bitstreams: 1 applsci_11_09774_v2.pdf: 185826 bytes, checksum: bb15ac9ae436d2df4e3fa3192a7b1f61 (MD5) Made available in DSpace on 2021-11-03T15:57:52Z (GMT). No. of bitstreams: 1 applsci_11_09774_v2.pdf: 185826 bytes, checksum: bb15ac9ae436d2df4e3fa3192a7b1f61 (MD5) Previous issue date: 2021-10-01 info:eu-repo/semantics/publishedVersion

Attacks using Uniform Resource Locators (URLs) and their JavaScript (JS) code content to perpetrate malicious activities on the Internet are rampant and continuously evolving. Methods such as blocklisting, client honeypots, domain reputation inspection, and heuristic and signature-based systems are used to detect these malicious activities. Recently, machine learning approaches have been proposed; however, challenges still exist. First, blocklist systems are easily evaded by new URLs and JS code content, obfuscation, fast-flux, cloaking, and URL shortening. Second, heuristic and signature-based systems do not generalize well to zero-day attacks. Third, the Domain Name System allows cybercriminals to easily migrate their malicious servers to hide their Internet protocol addresses behind domain names. Finally, crafting fully representative features is challenging, even for domain experts. This study proposes a feature selection and classification approach for malicious JS code content using Shapley additive explanations and tree ensemble methods. The JS code features are obtained from the Abstract Syntax Tree form of the JS code, sample JS attack codes, and association rule mining. The malicious and benign JS code datasets obtained from Hynek Petrak and the Majestic Million Service were used for performance evaluation. We compared the performance of the proposed method to those of other feature selection methods in the task of malicious JS code content detection. With a recall of 0.9989, our experimental results show that the proposed approach is a better prediction model.

Purpose: A deep learning technique was used to analyze the triceps surae muscle. The devised interpolation method was used to determine muscle’s volume and verify the usefulness of the method. Materials and Methods: Thirty-eight T1-weighted cross-sectional magnetic resonance images of the triceps of the lower leg were divided into three classes, i.e., gastrocnemius lateralis (GL), gastrocnemius medialis (GM), and soleus (SOL), and the regions of interest (ROIs) were manually defined. The supervised images were classified as per each patient. A total of 1199 images were prepared. Six different datasets separated patient-wise were prepared for K-fold cross-validation. A network model of the DeepLabv3+ was used for training. The images generated by the created model were divided as per each patient and classified into each muscle types. The model performance and the interpolation method were evaluated by calculating the Dice similarity coefficient (DSC) and error rates of the volume of the predicted and interpolated images, respectively. Results: The mean DSCs for the predicted images were >0.81 for GM and SOL and 0.71 for GL. The mean error rates for volume were approximately 11% for GL, SOL, and total error and 23% for GL. DSCs in the interpolated images were >0.8 for all muscles. The mean error rates of volume were

This study investigates social media trends and proposes a buzz tweet classification method to explore the factors causing the buzz phenomenon on Twitter. It is difficult to identify the causes of the buzz phenomenon based solely on texts posted on Twitter. It is expected that by limiting the tweets to those with attached images and using the characteristics of the images and the relationships between the text and images, a more detailed analysis than that of with text-only tweets can be conducted. Therefore, an analysis method was devised based on a multi-task neural network that uses both the features extracted from the image and text as input and the buzz class (buzz/non-buzz) and the number of “likes (favorites)” and “retweets (RTs)” as output. The predictions made using a single feature of the text and image were compared with the predictions using a combination of multiple features. The differences between buzz and non-buzz features were analyzed based on the cosine similarity between the text and the image. The buzz class was correctly identified with a correctness rate of approximately 80% for all combinations of image and text features, with the combination of BERT and VGG16 providing the highest correctness rate.

The deterioration of existing reinforced concrete (RC) structures is regarded as a problem worldwide. In Japan especially, the deterioration of RC dam structures has become severe. Many such dams meet design standards that were in place at the time of construction but do not meet the current seismic design standards, and appropriate seismic retrofitting is required. If the dam pier, which is an important part of the dam related to water storage, is damaged by an earthquake, the gate cannot be opened or closed, and the amount of water stored cannot be controlled. Therefore, the seismic retrofitting of dam piers is a top priority. However, various construction restrictions exist for dam piers, such as only the cross-section on the downstream side can be reinforced, and not on the upstream side where water is stored. Thus, it is difficult to apply the same reinforcement method that is applied to the piers of general road bridges. Therefore, in this study, we confirm the effectiveness of the SRS method (seismic retrofitting using cement mortar for shotcrete), which is suitable for partial reinforcement. Specifically, the dam piers of four types of existing dams were modeled using the three-dimensional finite element method, and a seismic response analysis was performed by inputting two types of seismic waveforms having different characteristics. As a result, the reinforcement effect of this method was verified according to structural characteristics. Furthermore, the effect of the reinforcement range on the reinforcement effect was clarified.

As a new method for evaluating single crystals and oligocrystals, pulsed neutron Bragg-dip transmission analysis/imaging method is being developed. In this study, a single Bragg-dip profile-fitting analysis method was newly developed, and applied for analyzing detailed inner information in a crystalline grain position-dependently. In the method, the spectrum profile of a single Bragg-dip is analyzed at each position over a grain. As a result, it is expected that changes in crystal orientation, mosaic spread angle and thickness of a perfect crystal can be evaluated from the wavelength, the width and the integrated intensity of the Bragg-dip, respectively. For confirming this effectiveness, the method was applied to experimental data of position-dependent Bragg-dip transmission spectra of a Si-steel plate consisting of oligocrystals. As a result, inner information of multiple crystalline grains could be visualized and evaluated. The small change in crystal orientation in a grain, about 0.4°, could be observed by imaging the Bragg-dip wavelengths. By imaging the Bragg-dip widths, both another grain and mosaic block in a grain were detected. Furthermore, imaging results of the integrated intensities of Bragg-dips were consistent with the results of Bragg-dip width imaging. These small crystallographic changes have not been observed and visualized by previous Bragg-dip analysis methods.

In this study, we analyzed the CO2 emissions of online music and video streaming services, as one of the digital contents, in Japan using life cycle assessment. As a system boundary of online music and video streaming, processes such as data center construction and server manufacturing, usage of communication networks and internet communication technology devices (personal computers (PCs) and smartphones), and disposal of data centers and servers were considered. Data were collected using statistical and online surveys, and CO2 emissions per 1 MB of communication volume were calculated. One of the results revealed that the lifecycle CO2 emissions of listening to online music using PCs and smartphones were 5.88 × 10−4 and 1.43 × 10−4 kg-CO2/MB, respectively. The overall CO2 emissions for domestic music and video streaming services in 2019 was 921 thousand t-CO2. Online video streaming accounted for 87.7% of the total emissions, which corresponded to approximately 0.23% of domestic CO2 emissions derived from electric power generation.