International Journal of Computer (IJC - Global Society of Scientific Research and Researchers, GSSRR)
Not a member yet
459 research outputs found
Sort by
Automating Homework Verification Through LLM Assistants
Full text linkThis article examines the automation of homework assessment through LLM assistants. A comprehensive architecture is proposed, comprising an Instruction Chains Generator for task decomposition, a Previous Action Description module for generating step summaries, an Action Prediction & Executor for planning and executing verification steps, and a Controllable Calibration component for refining outcomes. To ensure pedagogical soundness and increase reliability, the system integrates with Intelligent Tutoring System (ITS) logs and employs Retrieval-Augmented Generation (RAG) to mitigate model hallucinations. A prototype built on Llama 3 Instruct and the Ollama framework was evaluated in an online algebra course and the GSM8K benchmark (“problem + solution”). User studies with instructors confirmed the approach’s high explainability and the diagnostic value of its feedback. The results demonstrate the efficacy of a hybrid human + LLM workflow for automated homework grading. These findings will interest educational-technology researchers and AI developers aiming to embed next-generation language models in automated verification of student work, grounded in cognitive analysis and adaptive-learning methodologies. In addition to EdTech scholars and AI engineers, practicing educators and educational administrators focused on improving assessment quality and reducing grading workload through LLM assistants will find this work valuable
Machine learning applications for event routing in streaming systems
Full text linkThe paper provides a broad overview and classification of machine learning methods used to optimize routing in distributed streaming architectures. The aim of the study is to provide a detailed analysis of existing approaches: from classical reinforcement learning algorithms to modern deep neural networks, with an assessment of their potential in various operational scenarios and identification of key limitations. The methodological basis was a systematic review of publications dealing with intelligent routing, real-time data processing, and integration of ML solutions into system pipelines. Three main classes of algorithms were identified and considered: reinforcement learning methods (including DQN and actor-critic), deep networks (CNN, RNN and their hybrids), as well as ensemble and evolutionary techniques. The advantages and disadvantages of each class are analyzed in terms of key criteria — response time to flow changes, scalability in the number of nodes, and the ability to dynamically adapt. Special attention was paid to hybrid strategies that combine several models to increase the reliability and accuracy of recommendations on event transmission routes. In conclusion, the main conclusions about the current state of research are formulated and promising areas are outlined: the development of more robust architectures with explicable decision-making logic, as well as the integration of graph neural networks for modeling complex topologies of distributed systems. The presented results will be useful for engineers developing streaming platforms, big data analysis specialists, and research groups working on information channel optimization tasks
Federated Learning Approaches for Privacy-Preserving Conversational AI in Dental Informatics
Full text linkThis study investigates how federated learning can enable privacy-preserving conversational AI for dental clinics while complying with GDPR and HIPAA. The main objective is to determine whether clinics can automate patient communications, call handling, reminders, and triage, without transferring raw patient data beyond local systems. The methodology combines a narrative review and comparative analysis of recent healthcare federated-learning studies with a practical deployment blueprint tailored to dental workflows. Three training paradigms are contrasted, local, centralized, and federated, to summarize evidence on model accuracy and computational cost. A layered privacy stack is presented, including differential privacy, secure aggregation, and homomorphic encryption, with guidance on when each technique is most appropriate. An end-to-end workflow is described in which clinics train lightweight local adapters on anonymized speech and text, share only encrypted model updates, and receive an aggregated global model that improves across participating sites. Findings indicate that federated models achieve accuracy comparable to centralized baselines in representative tasks (e.g., segmentation and risk prediction) while keeping patient data local. Cryptographic protections introduce overhead but remain practical when applied selectively.
This study demonstrates that federated learning enables privacy-preserving conversational AI in dental informatics without compromising model performance. In practical terms, such systems can reduce missed appointments through targeted reminders, accelerate routing of urgent complaints, and support intern training with simulated dialogues, delivering the benefits of shared learning while maintaining strong confidentiality and regulatory compliance
A Review of Event-Driven Architecture Patterns Using Message Brokers in .NET
Full text linkThis article describes different event-driven architecture (EDA) patterns that use message brokers to implement distributed computing in .NET. This article also describes data consistency and state management techniques in a distributed system architecture. In response to the industry trend of moving from monolithic architecture to microservice architecture for scalability, agility, and resilience, this project seeks to provide a thorough basis for decision-making for the adoption of EDA patterns and message brokers (like Apache Kafka, RabbitMQ, Azure Service Bus) based on non-functional requirements and organizational maturity levels. The methodological foundation is based on a Systematic Literature Review (SLR), ensuring reproducibility, completeness, and methodological rigor of the analysis. The scientific novelty of this work lies in integrating three analytical dimensions: theoretical architectural patterns, the technological implementation of message brokers, and the practical aspects of applying them within .NET systems. The main findings emphasize that the choice of EDA patterns and broker technologies is not a search for an optimal, universal solution but a deliberate balancing of technical and organizational factors. Successful use of EDA requires maturity and experience in DevOps and observability, as well as strict adherence to idempotency principles. Therefore, the target audience for this article is software architects, developers, and researchers concerned with microservices and event-driven systems of all kinds built on .NET, as well as technical decision-makers responsible for enterprise adoption of EDA
Scaling MLOps in Pharma: Automating Model Deployment and Monitoring with DataRobot on AWS
Full text linkThis article discusses the MLOps scaling challenge amidst an increase in medical data and growing regulatory demands, which pharmaceutical companies are facing. On the AWS infrastructure, implementing DataRobot will make it possible to automate model deployment as well as monitoring and maintenance of models — without manual oversight; this traditional bottleneck usually creates another risk avenue for non-compliance with good manufacturing practice requirements. Such solution relevance is driven by an acute need for algorithm reproducibility and traceability under circumstances where pharmaceutical companies concurrently operate dozens of models. A lapse in validation or documentation can lead to clinical program delays and increased costs. The novelty here is a unified automated model registry plus deployment pipelines integrated with built-in drift/accuracy tracking and regulatorily significant audit systems. The major findings are that by automating the lifecycle, artificial intelligence stops being an artisanal collection of disconnected experiments and becomes a manageable production process. It shows how running DataRobot on AWS not only speeds up getting algorithms to the clinic but also makes sure strong FDA and GxP rules are followed with built-in version control, legally applicable report making, and data encryption. Such a strong setup where scientific change and rule order do not fight but instead help each other. The article will be of great use to drug makers, data engineers, MLOps workers, and rule managers who want to see the real use of auto ways for model work
Analyzing the Performance of ECLAT Algorithm for Large Datasets by Comparing K-means and Gaussian Mixture Model
Full text linkFrequent Itemset Mining (FIM) is a technique that transforms historical data into useful information by identifying beneficial patterns. The ECLAT method uses depth-first search to intersect the transaction ID sets with the corresponding kth item sets in order to calculate the support items. While searching for the best-selling products, ECLAT uses a lot of memory and processing time due to the enormous number of transaction ID sets. To overcome these problems, the clustering method combines with the ECLAT algorithm to retrieve the support items. Description elements 100,000 to 400,000 were used to retrieve the support items of the most popular selling goods. For the K-means clustering approach, the optimal value of k is 8 clusters according to the 0.59 silhouette value. For the Gaussian Mixture Model, the ideal value of k is 14 clusters based on a 0.59 silhouette score value between 100,000 and 400,000 data items. After clustering the same product items, the ECLAT algorithm retrieves the support items by applying a minimum support value of 0.00001 in this investigation. According to the experimental results, the Gaussian Mixture Model not only offers more flexibility for clustering the same items but also reduces the memory usage and execution times. The outcomes of this investigation indicate that the Gaussian Mixture Model provides more efficient enhancement of the performance of the ECLAT algorithm than the K-means algorithm
The Evolution of Test Automation: From Selenium to Playwright. A Comparison of Automation Tools: Selenium vs. Playwright vs. Cypress
Full text linkThis study examines the evolution of web application test automation tools, focusing on a comparative analysis of Selenium, Playwright, and Cypress. The research justifies the relevance of transitioning from the traditional Selenium-based approach to modern frameworks that offer higher performance and stability in the rapidly evolving landscape of web applications. The study follows a methodology that involves sequential execution of test scenarios of varying complexity, including a simple static site test, end-to-end testing in a production environment, and a comprehensive test suite evaluation. Key performance metrics such as average execution time, standard deviation, and coefficient of variation are assessed. The findings indicate that Playwright demonstrates the best performance for testing dynamic web applications, while Cypress, despite an initial slowdown in simple scenarios, becomes competitive when executing local test suites. The article provides practical recommendations for selecting an automation tool based on the characteristics of the tested applications and outlines future development prospects, including the integration of artificial intelligence technologies and the optimization of CI/CD processes. This research addresses an existing gap in the field and offers practical solutions to enhance the quality and efficiency of modern web application testing. The findings will be of interest to researchers in software engineering, quality assurance professionals, test system architects, and academic professionals seeking to integrate advanced methodologies into software development and testing processes
Leadership and Mentorship Models in Software Quality Management
Full text linkThis paper explores the synergy between leadership and mentorship models in driving improvements in software quality. Through a literature review, the study identifies that transformational and situational leadership—when combined with proactive mentorship practices—significantly enhance the effectiveness of quality assurance systems in IT projects. The structured Six Sigma methodology supports this process by reducing defects, optimizing development workflows, and enabling continuous improvement. The findings underscore that integrating managerial practices with the DMAIC framework serves as an effective means of cultivating a corporate culture centered on innovation and quality. Such integration is particularly vital for boosting competitiveness in the software development industry. The article offers valuable insights for both academic researchers and IT management professionals, as well as software quality experts aiming to embed modern leadership and mentorship theories into strategic management models to improve development and quality control outcomes. The paper’s analytical approach not only contrasts different leadership frameworks but also identifies optimal paths for implementing mentorship practices, thereby contributing to the advancement of management processes in the context of digital transformation
Artificial Intelligence in Design Systems: Advancing Scalable and Adaptive User Interface Frameworks
Full text linkArtificial intelligence has fundamentally reshaped design systems, establishing a new standard for modern interface development and digital product engineering. AI-powered frameworks drive efficiency by automating component creation, identifying design patterns, and monitoring accessibility compliance at scale. Through the application of neural networks, deep learning, and computer vision, these systems interpret user interactions and adjust interfaces dynamically. Evidence from testing healthcare and enterprise applications confirms that such systems consistently shorten development cycles, improve user satisfaction, and lower operational costs. Automation further enables design and engineering teams to shift focus toward higher-value innovation while ensuring uniformity across platforms. Machine learning enhances personalization, strengthens validation through automated testing, and ensures more reliable user experiences. Beyond measurable outcomes, AI integration improves accessibility compliance, decreases technical debt, and facilitates stronger collaboration across teams. Collectively, this evolution represents a decisive step forward in how organizations design, optimize, and sustain user interfaces in the digital era. This article will be especially useful for software engineers, UX designers, product managers, and academic researchers seeking to understand and implement AI-enabled design systems for scalable, adaptive, and efficient digital interfaces
Serial Communication Inter-IC Bus, Interface and Protocol Using BeagleBone
Full text linkTo communicate with peripherals the embedded systems mainly use the serial communication. Hence serial communication plays a vital role in designing embedded systems. The protocols used for serial communication are Universal Asynchronous Receiver Transmitter (UART), Serial Peripheral Interface (SPI), Universal Serial Bus (USB), Control Access Network (CAN) and Inter IC Protocol (I2C). The characteristics of serial communication protocol are high speed and low data loss, ensure the data transfer and simplifying the system level design. This paper provides an overview of I2C-Bus, I2C protocol and its interfacing. The paper also demonstrates the I2C protocol using BeagleBone