Asian Journal of Research in Computer Science
Not a member yet
792 research outputs found
Sort by
Leveraging AI for Environmental Sustainability and Climate Action
No full textThis research paper explores the intersection of artificial intelligence (AI), computer science, and environmental sustainability. As climate change accelerates, innovative technologies offer promising avenues for mitigation and adaptation. Through a review of existing literature and case studies, this paper highlights the transformative potential of AI in monitoring ecosystems, optimizing energy systems, and enhancing conservation efforts. The findings underscore the importance of interdisciplinary collaboration in harnessing these technologies for sustainable development. It concludes with recommendations for future research and policy directions
Advancing Information Governance in AI-Driven Cloud Ecosystem: Strategies for Enhancing Data Security and Meeting Regulatory Compliance
No full textThis study explores adaptive information governance models to address critical challenges in AI-driven cloud environments, focusing on enhancing data security and achieving regulatory compliance. Existing frameworks often fail to account for the complexities introduced by AI and cloud integration, leaving significant gaps in incident response, privacy protection, and governance practices. To bridge these gaps, this research evaluates governance components—Privacy-Enhancing Technologies (PETs), ethical oversight, and incident response metrics—through advanced quantitative methods, including Structural Equation Modeling (SEM), Cox Proportional Hazards Modeling, and Difference-in-Differences (DiD) analysis. Key findings highlight the significant influence of incident response metrics (β = 0.51, p < 0.001) and PET integration (β = 0.25, p = 0.001) on governance effectiveness, with model fit indices (RMSEA = 0.04, CFI = 0.96) confirming the robustness of the proposed framework. Industry-specific vulnerabilities were identified, with retail and technology sectors experiencing a 25% increased risk of incidents due to minimal security controls. The adoption of PETs, such as federated learning and homomorphic encryption, significantly improved privacy compliance and data utility, particularly in high-risk sectors. The study recommends the integration of advanced security controls and PETs to mitigate risks and improve compliance, especially in vulnerable industries like retail and technology. It also emphasizes the continuous optimization of AI-driven incident response protocols to reduce the impact of emerging threats. Furthermore, ethical oversight should be prioritized to ensure fairness, accountability, and public trust in AI applications within cloud ecosystems. These actionable strategies provide a roadmap for organizations to achieve secure and ethically governed cloud environments
Mitigating Artificial Intelligence Bias in Financial Systems: A Comparative Analysis of Debiasing Techniques
No full textBalancing fairness and predictive accuracy remains a key challenge in AI system development. This study investigates the origins of AI bias, how it happens in business processes, and the challenges it poses to ethical and transparent decision-making. Drawing on existing literature, the research explores the various types of biases—including cognitive, algorithmic, and representation biases—and their impact on AI systems in the BFSI sector. Furthermore, the study critically evaluates current debiasing techniques, such as pre-processing, fairness-aware models, and post-processing, highlighting their limitations in balancing fairness with predictive accuracy.
This study aims to advance the development of more equitable AI systems in the BFSI sector by proposing the FAIR-BIAS Framework. This framework provides a structured approach to detecting, mitigating, and monitoring biases in AI models. Key recommendations include implementing equalized odds as a fairness metric to ensure balanced outcomes across demographic groups, applying adversarial debiasing techniques during model training to minimize discriminatory effects, and conducting regular data audits to ensure long-term fairness.
The findings offer direct benefits for BFSI stakeholders. Businesses can enhance the reliability and ethical integrity of AI models by adopting fairness-aware risk assessments, which promote compliance and customer trust. Regulators can enforce accountability by mandating transparency measures, such as model explainability, and conducting periodic audits using fairness metrics like equalized odds. Policymakers can use the insights to create inclusive legislation which requires fairness testing and transparency in AI applications.
Future research could explore the long-term effectiveness of debiasing techniques across different industries, such as healthcare or public policy, by conducting longitudinal studies to assess how evolving datasets and models influence fairness outcomes.
It is critical for BFSI organizations to adopt these frameworks and techniques to foster a more inclusive and ethical future in financial services.
Enhancing Hospital Efficiency with Healthyline: A Digital Solution for Registration, Appointment Reminders, and Queue Management
No full textRegistration and queue management are essential for efficient healthcare services. However, the process has some potential drawbacks: long queues, long waiting times, difficulty in getting real-time information on queues, and missed doctor which negatively impact patient experience and hospital operations. So, this research addresses specific challenges in hospital operations, including inefficient queue management, high patient anxiety due to long waiting times, and frequent missed appointments. The study focuses on the development of the Healthyline application, which will provide online registration for patients, real-time queue information, and notification in regard to appointments with the doctor. The development process includes needs analysis, system design, development of web and mobile-based applications, testing, and maintenance. With the Healthyline application, online registration is possible, saving patients\u27 valuable time and reducing anxiety related to waiting times. The application also complements the operational efficiency of hospitals by advancing queue management and minimizing no-shows for appointments among patients, consequently enhancing the productivity of the hospital. Such an innovative solution is hypothetically bound to improve the lots of patients and healthcare providers in creating a more structured, efficient, and user-friendly healthcare environment. In a nutshell, Healthyline has been struggling to revolutionize the patient registration process, make queue management easy, and ensure good health for every stakeholder. This study is significant as it addresses critical issues in healthcare delivery, such as long queues and appointment management. By providing real-time updates and automatic reminders, the application not only enhances patient satisfaction but also improves hospital operational efficiency by reducing the number of no-show patients. While the study demonstrates significant improvements, it also acknowledges limitations such as the need for further evaluation of the application’s performance during peak hours and potential security concerns in handling patient data. It is important to note that the Healthyline application has not yet been implemented or deployed in a hospital setting; thus, the research is limited to the stages of development, testing, and maintenance
Cost-Efficient Deployment Strategies: A Comparative Analysis of Feature Flagging Services and Blue/Green Deployments
No full textAims: This study provides a detailed comparison of two leading feature flagging services, LaunchDarkly and ConfigCat, and also examines the cost-efficiency and operational implications of using Blue/Green deployment strategies. It seeks to aid stakeholders in understanding the trade-offs and benefits of each approach to make informed decisions based on project-specific needs, budget constraints, and developmental objectives.
Study Design: The study employs a comparative analysis framework, focusing on features, usability, scalability, architectural design, and the financial impacts of adopting feature flagging services versus Blue/Green deployment strategies.
Place and Duration of Study: The analysis was conducted over a period of two years, encompassing a broad range of software development environments and project scenarios to ensure comprehensive coverage and relevance.
Methodology: The methodology employed in this paper includes a detailed examination of LaunchDarkly and ConfigCat\u27s service offerings, an evaluation of Blue/Green deployment strategies, and an analysis of cost-efficiency of each approach. The study synthesizes information from product documentation, user feedback, and performance metrics, alongside interviews with industry experts and case studies from diverse software development projects.
Results: The results highlight the nuanced differences between feature flagging services in terms of scalability, ease of use, and the suitability for various project sizes. LaunchDarkly emerges as optimized for large-scale, complex projects due to its extensive feature set and scalability, while ConfigCat is favored for its simplicity and ease of use in smaller projects. The analysis also uncovers the cost benefits of feature flagging over Blue/Green deployments, emphasizing the savings on infrastructure and operational expenses while offering dynamic feature management capabilities.
Conclusion: The study concludes that the choice between feature flagging services like LaunchDarkly and ConfigCat, and the utilization of Blue/Green deployment strategies, should be guided by specific project requirements, financial constraints, and desired operational efficiency. Feature flagging services provide a cost-effective, flexible solution for dynamic feature management, whereas Blue/Green deployments offer a straightforward, though potentially more resource-intensive, approach to minimizing deployment risks. This comparative analysis aims to assist stakeholders in selecting the most appropriate deployment strategy to meet their development goals efficiently
NLP-Based Rule Learning from Legal Text for Question Answering
No full textLaw is a system containing rules and regulations that binds a people. Legal rules and regulations are usually expressed in domain specific terminologies which are presented in textual form. Its expression is not in machine understandable format for legal reasoning to infer new knowledge or determines if a course of action aligns with the law. Similarly, in order to conceive the rule layer of the semantic web vision in line with the W3C recommendation, in this paper, we present a rule learning technique for learning legal rules for legal question answering, where we learn rules from a collection of instance level triples to infer new rules which can be applied to facts to reason with to arrive at an answer. We explore the natural language processing tool to extract instance level triples from legal textual data and applied RUMIS tool on the extracted triples to produce nonmonotonic rules which are then translated and expressed in Semantic Web Rule Language for legal reasoning in answering legal questions. With the application of the mined rules with our handcrafted rules for legal reasoning, the system was able to answer six out of the thirty correct answers. The research output shows promising results with respect to rule learning for legal reasoning for question answering
Cross-Layer Energy Efficient (CLEE) Routing Algorithm for Mobile Ad-Hoc Networks
No full textAd-hoc routing algorithms in Wireless Sensor Networks (WSN) rely on nodes’ position awareness by regularly updating routing data of neighbouring nodes. Meanwhile, the transmission energy usage is not optimised as a result of this repeated updates and routing table deployment. Therefore, it is very critical to consider techniques of optimising or conserving energy in the design of wireless sensor networks (WSNs) in order to prolong the lifetime of the individual nodes. One of these techniques is the Cross-layer design which, is considered as an efficient method for addressing this challenge with WSNs. In this paper, we propose a Cross-Layer Energy Efficient (CLEE) routing algorithm to establish an optimal route from the source node to the destination node by selecting candidate nodes from neighbouring nodes based on the distance between these nodes and the rate of energy consumption by a possible candidate node. Then to select a designated node from the candidate nodes, the algorithm further computes the Signal Strength Quality (SQS) and the Link Lifetime (LL) as well as the Throughput (TH) rate of selected nodes. The proposed algorithm was simulated with a network size of 600 X 600 on Network Simulator 3 (NS3) in order to analyse its performance. An evaluation of the performance of the proposed CLEE protocol with existing similar protocols such as Ad-hoc On-Demand Distance Vector (AODV) and Dynamic Source Routing (DSR) reveals that CLEE outperformes AODV and DSR by conserving about 44% of available energy.
Security Vulnerabilities of Wlan Protocols: A Review
No full textThis study reviews a significant number of known attacks with the security methods that are currently available for wireless local area networks (WLANs). After a thorough review of more than a dozen literature on the security issues affecting WLANs, it was found that the architecture of the WLAN itself has a vulnerability that hackers can exploit, especially the wired equivalent privacy (WEP) protocol. WLAN attacks are a combination of human and technology behavior. Human network compromise is still a serious danger to wireless technology, although new encryption technologies are being developed to counteract threats, they are not enough to address the issues at hand because hackers are constantly coming up with inventive ways to breach networks and undermine the standards. Therefore, it is advised to integrate more sophisticated and pervasive multidimensional AI algorithms at different wireless protocol layers as well as through certified ethical hacking experimentation standards in order to improve detection capabilities, identify unauthorized entry through their behaviors, and safeguard wireless connections and data. Also, Cyber security professionals should take note of this report and be proactive in their search for a sophisticated solution to the current issue
Leveraging AI and Machine Learning for the Protection of Critical National Infrastructure
No full textNo nation can exist or survive without critical infrastructure (CI), which is why a nation’s growth, development, welling, standard of living, possessions, and even governance are weighed by the kind of CI obtained therein. There are growing concerns about the need and how to protect CI from cyber threats in the 21st century era of digitalization. This descriptive survey research aims at showing how artificial intelligence (AI) and machine learning (ML) can be leveraged for the protection of critical national infrastructure (CNI). The study relies on secondary data, which are subjected to thematic systematic review. Interpretive and descriptive analytic techniques are used. The analysis shows that leveraging AI and ML for the protection can yield huge results, as they optimize detection of and response to threats, facilitate efficient physical maintenance, optimally evaluate and manage risks, increase awareness, and simulate and train human employees in the CNI sector. The study concludes that these cutting edge technologies have more capacities and opportunities for the protection of CNI from cyber threats than other non-technological and less advanced technological mechanisms. It calls on stakeholders, especially national governments and authorities of the organizations involved in CNI, to make concerted efforts to surmount the challenges of AI and ML adoption and ensure significant protection of CNI across nations of the globe. Doing so would pave way for extensive practical usage of AI and ML for the protection of CNI
Object Detection Algorithms Based on Deep Learning: A Review
No full textWith the continuous development of deep learning, object detection algorithms based on deep learning have made significant progress in the field of computer vision, widely applied in areas such as autonomous driving, industrial inspection, agriculture, transportation, and medicine. Traditional object detection algorithms face issues such as low detection efficiency and poor robustness. However, deep learning-based object detection algorithms significantly enhance detection accuracy and generalization by learning low-level and high-level image features. This article first introduces traditional object detection algorithms and their existing problems, then elaborates on the main processes, innovations, advantages, disadvantages, and experimental results on datasets of deep learning-based object detection algorithms. It focuses on the development of Two-Stage and One-Stage object detection algorithms, and provides an outlook on the future development of object detection algorithms, discussing challenges such as the coordination of detection speed and accuracy, difficulties in detecting small objects, real-time detection tasks, and multi-modal fusion applications, and proposes possible future directions