Faculty of Computer Science and Engineering

Permanent URI for this communityhttps://repository.ukim.mk/handle/20.500.12188/5

The Faculty of Computer Science and Engineering (FCSE) within UKIM is the largest and most prestigious faculty in the field of computer science and technologies in Macedonia, and among the largest faculties in that field in the region. The FCSE teaching staff consists of 50 professors and 30 associates. These include many “best in field” personnel, such as the most referenced scientists in Macedonia and the most influential professors in the ICT industry in the Republic of Macedonia.

Browse

Search Results

Now showing 1 - 10 of 61

Some of the metrics are blocked by your
consent settings
Optimizing document retrieval using massive text embeddings and LLM prompt engineering
(Springer Science and Business Media LLC, 2026-04-14)
Mitrov, Goran
;
Stanoev, Boris
;
;
;
Kampel, Martin
Background The rapid expansion of digital data poses a unique challenge for retrieving relevant and insightful information efficiently. In particular, the increasing volume of scientific publications has made literature reviews time-consuming. The emergence of large language models (LLMs) offers new opportunities to streamline this process. Methods This paper explores the use of generative artificial intelligence (GenAI) for query reformulation and evaluates the performance of nine massive text embedding models, varying in size and fine-tuning strategies, in the context of document retrieval. We apply multiple prompt engineering techniques to evaluate the ability of LLMs to generate effective queries, comparing them with human-crafted queries. These are used to retrieve documents utilizing nine embedding models. The evaluation is across five datasets using metrics such as recall, average precision, and rank-based measures. Results Results show that embedding models fine-tuned for semantic similarity consistently outperform general-purpose models, with UAE Large proving most robust across diverse domains. Furthermore, queries generated using zero-shot and few-shot prompting techniques often surpass the performance of human-formulated queries. Conclusion These findings highlight the value of integrating LLMs and massive text embeddings to reduce manual effort in literature reviews. GenAI provides a reliable starting point for query formulation, with human input reserved for refinement when needed.
Some of the metrics are blocked by your
consent settings
An Empirical Study of Knowledge Graph-Enhanced RAG for Information Security Compliance
(MDPI AG, 2026-04-20)
Jovanovski, Dimitar
;
Stojcheva, Marija
;
Dodevska, Mila
;
;
Information security compliance has become critical for organizations worldwide, with the ISO/IEC 27000 family serving as the most widely adopted framework for establishing information security management systems. Despite their global acceptance, these standards present significant interpretation challenges due to their formal language, abstract structure, and extensive cross-referencing across 97 documents. Traditional retrieval-augmented generation (RAG) systems, which rely on independent text chunking and dense vector retrieval, prove inadequate for such highly interconnected regulatory materials, often fragmenting contextual relationships and reducing accuracy. This study introduces a privacy-preserving RAG framework that integrates LightRAG, a knowledge graph-based retrieval system, with locally hosted open-source language models. Unlike chunk-based RAG systems that treat document segments independently, the system in this study constructs a semantic knowledge graph that explicitly models relationships between clauses through typed edges representing cross-references, semantic similarity, and hierarchical dependencies. To enable rigorous evaluation, we developed a curated benchmark dataset of 222 multiple-choice questions with authoritative ground-truth answers, systematically constructed from official ISO standards, certification preparation materials, and academic sources. Through systematic evaluation on this benchmark, we show that knowledge graph-based retrieval achieves higher accuracy than chunk-based RAG and non-retrieval LLM baselines within the evaluated setup. The analysis indicates that embedding model quality is strongly associated with system performance, that hybrid retrieval modes combining local and global graph traversal tend to yield better accuracy, and that mid-sized open-source models paired with strong retrievers can approach the performance of larger proprietary systems. The best configuration achieves 90.54% accuracy, demonstrating the promising effectiveness of graph-structured retrieval for multiple-choice regulatory questions.
Some of the metrics are blocked by your
consent settings
Benchmarking OpenAI's APIs and Large Language Models for Repeatable, Efficient Question Answering Across Multiple Documents
(Polish Information Processing Society, 2024-10-23)
Filipovska, Elena
;
Mladenovska, Ana
;
Bajrami, Merxhan
;
Dobreva, Jovana
;
Hillman, Vellislava
Some of the metrics are blocked by your
consent settings
Colonoscopy image analysis for polyp detection: A systematic review of existing approaches and opportunities
(Elsevier BV, 2025)
Albuquerque, Carlos
;
Neves, Paulo Alexandre
;
Godinho, António
;
;
Some of the metrics are blocked by your
consent settings
A low-cost device-based data approach to Eight Hop Test
(Elsevier BV, 2025)
Pimenta, Luís
;
Coelho, Paulo Jorge
;
Gonçalves, Norberto Jorge
;
Lousado, José Paulo
;
Albuquerque, Carlos
Some of the metrics are blocked by your
consent settings
Weather-Aware Deep Reinforcement Learning for Predictive Modeling of Household Energy Dynamics
(Istanbul University, 2026-01-30)
Bajrami, Enes
;
This study proposes a weather-aware deep reinforcement learning (DRL) framework for predictive modelling of household energy dynamics. Using a 14-month high-resolution dataset from a residence in Northeast Mexico, the framework integrates detailed meteorological attributes and next-day forecasts to enhance prediction accuracy. Four DRL algorithms were implemented and evaluated for their performance in forecasting household energy consumption: Proximal Policy Optimisation (PPO), Soft Actor-Critic (SAC), Deep Deterministic Policy Gradient (DDPG), and Asynchronous Advantage Actor-Critic (A3C). Exploratory data analysis revealed significant seasonal trends and variability in energy usage patterns. Results show that DDPG and SAC outperform PPO and A3C, achieving the lowest root mean square error (RMSE) and mean absolute error (MAE), with DDPG recording 0.0011 RMSE and 0.0009 MAE. The framework was tested on moderately equipped hardware, demonstrating the practical feasibility of DRL-based energy forecasting systems. This work contributes original visualisations and comparative insights, advancing smart energy management solutions.
Some of the metrics are blocked by your
consent settings
Modelling and quantifying numerical integration errors in deep reinforcement learning for propulsion dynamics
(Elsevier BV, 2026-10)
Bajrami, Enes
;
Bajrami, Ensar
;
This study investigates how numerical integration accuracy influences the training dynamics and control performance of deep reinforcement learning controllers applied to propulsion system simulations. The propulsion dynamics are represented by a continuous second-order thrust-driven model that is discretised using four numerical integration configurations: Euler (coarse, medium, and fine time steps) and Runge-Kutta fourth order (RK4). Three widely used model-free reinforcement learning algorithms, Proximal Policy Optimization (PPO), Soft Actor-Critic (SAC), and Twin Delayed Deep Deterministic Policy Gradient (TD3), are evaluated together with a linear proportional-derivative baseline controller. A large experimental campaign comprising more than 50,000 simulated episodes was conducted across three training phases to quantify the influence of discretisation accuracy on reward convergence, trajectory stability, and control energy. The results demonstrate that numerical integration fidelity significantly shapes the optimisation landscape experienced by reinforcement learning agents. Under coarse Euler discretisation, PPO exhibits unstable learning behaviour and large oscillatory trajectories, while SAC maintains improved robustness but still shows sensitivity to large time steps. TD3 demonstrates the highest tolerance to discretisation error, maintaining stable closed-loop dynamics even under coarse integration. Higher-accuracy numerical schemes substantially improve learning efficiency. The RK4 configuration produces smoother trajectories, reduced control energy, and faster convergence across all reinforcement learning algorithms. Quantitative analysis of trajectory stability, integrated error metrics, and reward statistics confirms that discretisation error directly propagates through the learning process and alters the resulting control policies. These findings provide new empirical evidence that numerical integration fidelity is a critical design factor for reinforcement learning environments involving dynamical systems. The study highlights the necessity of carefully selecting integration schemes when training reinforcement learning controllers for propulsion dynamics and other physics-based control applications.
Some of the metrics are blocked by your
consent settings
AI Act Compliance Within the MyHealth@EU Framework: Tutorial
(JMIR Publications Inc., 2025-11-10)
;
;
Dobreva, Jovana
;
Bukovec, Djansel
;
Gjorgjioski, Blagojche
Some of the metrics are blocked by your
consent settings
AI in Software Testing: Revolutionizing Quality Assurance
(IEEE, 2024-11-26)
Trifunova, Andrea
;
;
;
Artificial intelligence (AI) is an area of tremendous potential, especially in the software testing domain, where it has changed the dynamics of the process, storms in efficiency, accuracy, and flexibility in a given SDLC. This paper presents findings from recent investigations of AI in the testing and quality assurance focusing on its transformational potential. Particular attention is paid to such issues as automation of testing processes through AI, testing process enhancement, and possible changes in software engineering due to AI implementation. In this paper, various research perspectives have been integrated to reveal the effectiveness of AI in enhancing the perceived quality assurance processes, improving product quality, and adopting principles of agile methodology in today's software development.
Some of the metrics are blocked by your
consent settings
Virtual reality as a learning tool: Evaluating the use and effectiveness of simulation laboratories in educational settings
(Elsevier BV, 2025-01-01)
Dodevska, Mila
;
;
;
Branco, Frederico
;
Virtual Reality Laboratories (VRLs) are essential progress towards educational solutions that would allow students to learn the sometimes complex experimental processes without spending physical resources. The resources used in some experiments often require significant procurement effort, significantly impacting the environment where an experiment would be conducted. The main focus of this research is to review the possibility of using VR solutions as educational tools and innovative approaches to scientific knowledge production. This paper systematically reviews relevant publications in this area while covering several existing virtual lab solutions. The analysis shows that VRL tools can be effectively used for educational purposes, allow access to lab resources for people with disabilities, and could be used to reach the desired learning outcomes. A gap exists between natural laboratories and VRLs, especially in the collaborative aspect of the laboratory exercises. However, there is significant ongoing research on this topic and a high potential for development.

Faculty of Computer Science and Engineering

Browse

Filters

Settings

Sort By

Results per page

Search Results