Moje publikacje naukowe
A Hybrid LLM-Based Architecture for Reliable Standardization of Heterogeneous Grant Information
Business Information Systems
Abstract
Public grants are an important instrument for financing socio-economic development, innovation, and public policy objectives. However, identifying and comparing key information on available grants remains difficult due to the fragmented, heterogeneous, and largely unstructured manner in which grant information is published by entities administrating grants. Recent advances in LLMs enable new approaches to processing unstructured textual data and generating standardized representations that summarize and enhance key information and thus facilitate comparison. Nevertheless, their direct application in domains requiring high informational reliability is constrained by issues such as hallucination and limited reproducibility. This paper examines the practical applicability and limitations of LLMs for standardizing grant information originating from heterogeneous, human-created sources. We propose a hybrid algorithmic architecture that bifurcates the standardization process into two complementary pathways. Descriptive textual fields are generated using a tightly controlled, parameter-constrained LLM pipeline, while hard factual data are extracted using predetermined, rule-based methods. This design restricts the scope of generative models to mitigate hallucination risks while preserving flexibility in processing unstructured content. The proposed solution was implemented as a web-based system and successfully evaluated under real-world conditions. The results demonstrate that a constrained and selective application of LLMs can support effective standardization of heterogeneous information while mitigating the risks associated with hallucination and inconsistency. The proposed approach is not limited to the grant domain but is of a general nature and can be transferred to other business areas characterized by fragmented, unstructured, and inconsistently published information.
Analyzing Novice Programming Behavior in Educational Game Environment: A Custom Language Study with Semantic Adaptation
Proceedings of the 18th International Conference on Computer Supported Education
Abstract
Programming education involves balancing syntactic complexity with conceptual understanding, particularly for novice learners. This paper investigates how a simplified custom programming grammar supports early concept acquisition in a game-based learning environment. The proposed environment combines an educational programming language with mechanisms for learner modelling and adaptive support, enabling systematic observation of learner behaviour during progressively more complex programming tasks. To examine the effects of syntactic simplification, we conducted a two-part evaluation consisting of Halstead complexity analysis and a user study focused on learning progression, problem-solving behaviour, and engagement. The results show that simplified syntax supports early task completion and initial concept acquisition, while increasing task complexity leads to shifts in solution strategies, cognitive effort distribution, and completion rates. These findings provide insight into how g rammar design influences early programming learning and suggest that effective programming education should balance syntactic accessibility with gradual increases in conceptual difficulty.
AI-Driven Video Avatar for Academic Support
Business Information Systems
Abstract
This paper presents the implementation and initial evaluation of an AI-powered interactive video avatar designed to enhance student consultations in a university course. As educational institutions explore new ways to provide accessible and flexible academic support, we investigate the potential of AI avatars in consultation scenarios. The implementation combines video synthesis with natural language processing to create an interactive professor avatar trained on content from actual lectures. We evaluated the system with students through structured tasks focused on technical knowledge delivery, explanation capabilities, interaction and context management, and handling ambiguous queries. The assessment covered both educational effectiveness in sustaining meaningful dialogue and technical performance - including audiovisual synchronization and animation quality. Results demonstrate the avatar’s effectiveness in conveying technical content and maintaining educational context during consultations. Technical performance was particularly strong in speech quality and audiovisual synchronization, demonstrating robustness even when addressing technical queries on multimedia topics. These findings highlight the potential of AI avatars to enhance student consultations in technical academic subjects.
Integrating Generative AI into Narrative-Based Educational Games for Moral Development: A Conceptual Approach
Avant
Abstract
This paper presents a conceptual approach for integrating artificial intelligence (AI) into the mechanics of narrative-based educational games. Drawing on theoretical analysis and a comprehensive literature review, the proposed model emphasizes the use of generative artificial intelligence (GenAI) to personalize game worlds, adapt storylines dynamically, and simulate the consequences of players’ moral decisions. Rather than offering static scenarios, the system enables real-time narrative adaptation based on decision patterns and player profiles, promoting deeper engagement and ethical reflection. The core contribution lies in demonstrating how AI can support the development of critical thinking and moral sensitivity through interactive storytelling. The proposed solutions may inform the design of educational games and training environments that foster moral development in a socially responsible and pedagogically grounded manner.
Advancing STEM Education in Primary Schools with an Integrated System of 3D Games
Proceedings of the 28th International ACM Conference on 3D Web Technology
Abstract
This paper introduces a method to enhance STEM education in primary schools by integrating 3D educational games with traditional teaching tools. Our approach combines two fundamental aspects: technical, which incorporates multi-platform educational games of various genres, and social, fostering interactions between teachers, students, and software developers. This model encourages adaptive learning tailored to student needs and promotes continuous student engagement. By simplifying complex STEM concepts through gamification, our method holds promising potential to improve the quality and effectiveness of STEM education in primary schools.