Software defect prediction aims to identify faults in code before they cause failures. It helps developers locate problematic modules early in the development cycle. Early detection reduces the cost of fixing defects and prevents them from affecting users. Just-in-Time defect prediction is the concept of detecting buggy commits early on in the development stage rather than in the release. While recent research focuses on improving the accuracy of pre-diction faulty commits, fewer studies explore how to do that task with minimal effort. In this study, we propose an effort-aware JIT defect prediction pipeline that utilizes code pre-trained models finetuned with Low Rank Adaptation (LoRA) and a feedback loop mechanism. The aim is to improve fault detection while reducing effort. Our experiments evaluate four different transformer-based models: CodeBERT, JavaBERT, UniXcoder, and RoBERTa on ApacheJIT. Our results illustrate CodeBERT and UniXcoder are benefiting the most from the feedback loop with the highest gain in F1 score, where CodeBERT highest gain is 0.14 and UniXcoder is 0.02. RoBERTa performed the best across experiments with an F1 score of 0.73. However, Popt metric remains static with limited gain that reached maximum 0.02, suggesting that further experiments are needed to improve performance.

Liver cancer is a complex disease responsible for a high number of deaths across the globe each year, making automated solutions for liver cancer classification urgent. The most common form of liver cancer is hepatocellular carcinoma (HCC), accounting for over 90 % of liver cancer cases. There is a distinct lack of publicly available HCC datasets utilizing genomic data, which is necessary for training artificial intelligence (AI) models for automated HCC classification. This study proposes constructing a multi-stage HCC dataset using XGBoost and Semi-Supervised learning on three separate datasets of genomic biomarkers, utilizing their existing labels in the Semi-Supervised learning process to label the proposed dataset. The proposed dataset consists of 770 patient samples in total, categorized into five classes that represent normal tissue alongside different stages of HCC. Each sample in the dataset consists of$\mathbf{1 1, 1 5 0}$different gene expression levels. The XGBoost model demonstrated a final classification accuracy of 96.5 % during the Semi-Supervised learning process.

This research delves into the impact of AI code generation tools, focusing on ChatGPT, programmers, and developers. Through interviews with 32 participants from diverse programming backgrounds, the study explores various aspects such as experience levels, language preferences, task complexities, completion speeds, prompt usage, output quality, ease of use, benefits, challenges, user scenarios, and feedback. The findings highlight a notable engagement with AI tools, particularly among those with several months of experience. Python emerges as the most used language alongside ChatGPT. While AI tools excel in more straightforward tasks and offer quick responses, challenges arise in complex coding scenarios where correctness may be compromised. Frequent prompts are favored for accuracy, yet output quality varies based on query clarity. Overall, AI tools are perceived as user-friendly, though complexities exist. Users appreciate their efficiency but caution against overreliance, especially for beginners. Recommendations include continuous enhancement for handling complex tasks. This study provides valuable insights for integrating AI tools into programming workflows responsibly.

A deepfake is a fast attack technique that has evolved rapidly over the past several years. It only takes one person to synthesize thousands of photorealistic images in a few hours or to manipulate a large number of videos. The creation of fake faces through image tampering has been identified as abusive media and could result in major ethical, legal, or political consequences. Existing known methods, such as Generative Adversarial Network (GAN), have simplified the synthetization of such images and can be used for detection purposes. However, current solutions and existing methods remain vulnerable and cannot stop deepfakes from spreading. In this paper, two novel solutions are proposed, named ColorDense and LightDense. They both use the DenseNet network backbone to determine fake face images from real face images. The two proposed models incorporate luminance and color signals from multi-color spaces to spot manipulated images obtained from the most recent deepfake datasets. Chrominance components of face images are considered to give a deep representation, while luminance depends on measuring the perceived brightness levels of the image. The system extracts the important features using the DenseNet network. Afterwards, it then passes the feature vector to a classification system based on neuronal techniques such as Multilayer Perceptron (MLP), Random Forest (RF), Support Vector Machine (SVM), and Radial Basis Function (RBF) as deepfake image detectors. Several experiments were conducted in this paper and the proposed system achieved over 98% accuracy in classifying deepfake images.

This paper encapsulates the fundamental operation of a Quantum Radar and demonstrates how entanglement can improve a radar’s signal-to-noise ratio. It focuses on the evolution of Quantum Radar technology by examining various types and bringing to light their unique characteristics. The paper also discusses the optimal radar type for various situations. Moreover, it thoroughly explores the properties of the Quantum Radar cross-section and compares it with the classical radar. Based on the research, finding covert targets and navigating through jammed environments are two key advantages of Quantum Radar. Finally, the paper discusses the limitations of each Quantum Radar type, along with future research directions.

Aircraft detection has become a very crucial aspect in the aviation industry due to its vital role in identifying unauthorized aircraft trespassing into restricted areas. Due to their significant role in decision-making, these detection systems must have a high level of accuracy. In this paper, we address deep learning feature extraction and aircraft detection methods. This study aims to explore deep learning-based feature extraction techniques and airplane identification techniques. The applications examined stop unauthorized aircraft from entering restricted airspace by utilizing a range of deep learning models. The study results and basic information on feature types, image types, and detection techniques are presented in this paper.

In the field of artificial intelligence (AI), deep learning (DL) has become an essential methodology with many uses, particularly in the aviation sector. The efficiency and effectiveness of tasks connected to aircraft recognition and classification are considerably improved by DL. The goal of this study is to examine and deal with problems related to efficiency and accuracy in DL models for identifying and classifying aircrafts. In this article, we suggest a solution that utilizes a customized convolutional neural network (CNN) model that our team has created and assess its performance in comparison to other deep learning (DL) models like VGG16 and ResNet50. Accuracy, recall, precision, and f1-score are among the metrics used for evaluation. The outcomes of this study demonstrate significant gains and offer insightful information for future endeavors in the aviation field.

Deploying drones in different real-life world applications can enhance user experience and maximize system benefits. The “Drone Box” project launched by Dubai aims to reduce the response time to traffic accidents and crimes; therefore, identifying optimal locations for these boxes is crucial. We propose a hybrid AI-GIS-based system to localize the drone boxes across Dubai strategically. We formulate the problem as a P-center Facility Location Problem (FLP) and construct a suitable objective function accordingly. Since the dataset needed was not provided, we designed a random dataset generator for testing purposes. We selected the Genetic Algorithm (GA) for benchmarking and investigated the performance of Artificial Bee Colony (ABC) algorithms due to their relevance to FLP. Geographic Information System (GIS) functions were employed to construct the random dataset generator and implement the ABC algorithm, utilizing the additional functionalities GIS offers for handling geographical data. After testing both algorithms, we deduced that the GA without GIS function integration consistently achieved a higher fitness value in a remarkable time, demonstrating its superiority over the ABC Algorithm.

The built environment sector’s growth has resulted in a rise in energy demands, which current energy production techniques cannot fulfill sustainably. Therefore, accurate energy management and forecasting tools are necessary to discover cleaner and more efficient energy production methods. This study proposes a Long-Short Term Memory (LSTM) model to forecast energy usage based on 24-hour periods of historical data. We evaluated the proposed model using the Building Data Genome dataset to determine its effectiveness in solving the energy consumption forecasting problem and observed superior performance compared to the Support Vector Regression (SVR) method. The LSTM model achieved an RSME score of 0.0710, an MAE score of 0.0506, and a MAPE score of 5.0557. This study provides additional experimentation to the existing literature and presents a consistent comparison between deep learning and machine learning techniques for energy consumption forecasting.

Due to the rapid population growth and economic development, water management became a necessity to assure sustainability. Analyzing water consumption patterns enables decision makers to better manage resources to meet the current demands without compromising future needs. This research paper focuses on investigating water consumption patterns through time series clustering in Dubai, United Arab Emirates, as one of the major water-stressed cities. Agglomerative hierarchical was applied to cluster the consumption patterns into multiple groups based on behavioral similarities. The consumption behavior of each cluster is analyzed based on the residential, commercial, and industrial sectors from 2017 to 2020. The study resulted in classifying the datasets into five clusters, in which the residential sector had the highest consumption, followed by the commercial and industrial sectors. Moreover, the majority of clusters reported high water consumption in 2020, except for clusters with relatively low number of accounts. In addition, the effect of data clustering was investigated in order to improve water demand forecasting. Based on the results, time series clustering increased the accuracy of predictions, in which artificial neural network and random forest models obtained the highest accuracy specially in clusters with high number of observations. It was found that the forecasting performance was highly correlated to the number of communities in each cluster.

Buildings are the most energy-intensive in cities, where they are designed to meet the needs of employees, customers and machines. The energy consumption of these buildings varies throughout the day. Their highest energy consumption is often during working hours, when employees and customers consume more energy in the building. Consumption is reduced at times when no one is in the building. However, it is difficult to manually disconnect power from the machines and air conditioning. Energy consumption is a major cause of pollution. It can affect and damage the environment, impacting society as a whole. As a result, we need to find ways to reduce energy consumption, and one way to do this is to shift to systems that are automatically managed, without any inter-human or human-to-machine interaction. Energy consumption can be decreased through the use of building energy management systems (BEMS) and the Internet of Things (IoT). BEMS plays an important role in improving efficiency, economics, reliability and energy conservation for distribution systems. IoT is a concept employed in the functioning of everyday objects, from industrial machines to wearable devices, using builtin sensors to collect data. Data is recorded via the Internet, and sensors are used to perform tasks such as automatically adjusting heat and lighting, or alerting service personnel in case of emergency failure. IoT is the future of technology, and will make our lives more efficient in all sectors. Our contribution in this research paper is to review various methods that use the Internet of Things to reduce building energy consumption by comparing and analyzing the advantages and disadvantages of existing methods and proposed plans. This will give us a comprehensive view of how IoT can be implemented in an efficient way.

Blockchain is a key technology that has the potential to decentralize the way we store, share, and manage information and data. One of the more recent blockchain platforms that has emerged is Hyperledger Fabric, an open source, permissioned blockchain that was introduced by IBM, first as Hyperledger Fabric v0.6, and then more recently, in 2017, IBM released Hyperledger Fabric v1.0. Although there are many blockchain platforms, there is no clear methodology for evaluating and assessing the different blockchain platforms in terms of their various aspects, such as performance, security, and scalability. In addition, the new version of Hyperledger Fabric was never evaluated against any other blockchain platform. In this paper, we will first conduct a performance analysis of the two versions of Hyperledger Fabric, v0.6 and v1.0. The performance evaluation of the two platforms will be assessed in terms of execution time, latency, and throughput, by varying the workload in each platform up to 10,000 transactions. Second, we will analyze the scalability of the two platforms by varying the number of nodes up to 20 nodes in each platform. Overall, the performance analysis results across all evaluation metrics, scalability, throughput, execution time, and latency, demonstrate that Hyperledger Fabric v1.0 consistently outperforms Hyperledger Fabric v0.6. However, Hyperledger Fabric v1.0 platform performance did not reach the performance level in current traditional database systems under high workload scenarios.

Security and the privacy features concerning e-banking needs to be improved rapidly to continue its growing. It is really difficult to ensure enough adequate security by using the conventional algorithms for a long time period, due to recent advances such as high progress in cryptanalysis techniques, improvement of computing skills and continuous hacking trials. This paper refers important issues regarding how to enhance the transition to more secure cryptographic and encryption algorithms in the financial sector. This paper recommends that adopting and implementing open source applications following international standards can be considered as a good replacement to the conventional algorithms to offer more enhancement security techniques and highest performance encryption algorithms for e-banking transaction services. We proposed a modified algorithm for AES, in which substitute byte, shift row will remain as in the original AES while mix column operation is replaced by 128 permutation operation followed by add round key operation. Comparative study with traditional encryption algorithms is shown the superiority of the modified algorithm and its high ability to overcome the problem of computational overhead. We additionally suggested another level of e-banking security services using Confidence Building Metric (CBM). The CBMs are computed based on certain parameters and can be implemented on any platform at the client side.

OBJECTIVE - The objective of this paper is to describe a case study where Bayesian Networks (BNs) were used to construct an expert-based Web effort model. METHOD - We built a single-company BN model solely elicited from expert knowledge, where the domain expert was an experienced Web project manager from a small Web company in Auckland, New Zealand. This model was validated using data from 22 past finished Web projects. RESULTS - The BN model has to date been successfully used to estimate effort for numerous Web projects. CONCLUSIONS - Our results suggest that, at least for the Web Company that participated in this case study, the use of a model that allows the representation of uncertainty, inherent in effort estimation, can outperform expert-based estimates. Another nine companies have also benefited from using Bayesian Networks, with very promising results.

Although the College of Information Technology (CIT) at Zayed University follows the ACM guidelines for Information Security curricula, its graduates are not able to fully meet employers' requirements. In this paper, we illustrate a new approach for teaching and engaging students in the context of a real experience related to the Information Security field using ISO 27001. Ten IT students at the college were supervised throughout their capstone projects, in which they investigated the use of ISO standards related to IT in the UAE. They expressed a great deal of satisfaction with their projects, and, created five case studies. Three of these are related to ISO 27001 implementation. In addition, three of the students were hired to work in this area after graduation. Our results reveal the importance of integrating international standards into the curricula of educational institutions.

This research adopts a new scenario-based black box testing methodology for testing web applications. It combines a black box testing strategy with the functions (scenarios) measured by the COSMIC-FFP measurement procedure (ISO/IEC 19761 standard) to produce an optimal set of test cases. This testing approach shows its applicability during all the development phases. Moreover, it can be applied during the early development phase once the specifications have been documented as well as after the development phase where we don't have the access to the code. This paper also considers the use of a functional complexity measure for assigning priorities to the generated test cases. Finally, those concepts have been applied on part of Online Banking System as a case study.

This paper presents an overview of quantitative analysis techniques for software quality and their applicability during the software development life cycle (SDLC). This includes the Seven Basic Tools of Quality, Statistical Process Control, and Six Sigma, and it highlights how these techniques can be used for managing and controlling the quality of software during specification, design, implementation, testing, and maintenance. We verify whether or not these techniques, which are generally accepted for most projects, most of the time, and have value that is recognized by the peer community, have indeed been included in the SWEBOK Guide.

Software component technology has a substantial impact on modern IT evolution. The benefits of this technology, such as reusability, complexity management, time and effort reduction, and increased productivity, have been key drivers of its adoption by industry. One of the main issues in building component-based systems is the reliability of the composed functionality of the assembled components. This paper proposes a reliability assessment model based on the architectural configuration of a component-based system and the reliability of the individual components, which is usage- or testing-independent. The goal of this research is to improve the reliability assessment process for large software component-based systems over time, and to compare alternative component-based system design solutions prior to implementation. The novelty of the proposed reliability assessment model lies in the evaluation of the component reliability from its behavior specifications, and of the system reliability from its topology; the reliability assessment is performed in the context of the implementation-independent ISO/IEC 19761:2003 International Standard on the COSMIC method chosen to provide the component's behavior specifications. In essence, each component of the system is modeled by a discrete time Markov chain behavior based on its behavior specifications with extended-state machines. Then, a probabilistic analysis by means of Markov chains is performed to analyze any uncertainty in the component's behavior. Our hypothesis states that the less uncertainty there is in the component's behavior, the greater the reliability of the component. The system reliability assessment is derived from a typical component-based system architecture with composite reliability structures, which may include the composition of the serial reliability structures, the parallel reliability structures and the p-out-of-n reliability structures. The approach of assessing component-based system reliability in the COSMIC context is illustrated with the railroad crossing case study.