ABSTRACT. Artificial Intelligence (AI) has transformed the way we interact with data and information. The growing volume of data and knowledge has made knowledge visualization (KV) important tool for understanding and analyzing complex datasets. AI has the potential to significantly improve KV by automating tasks such as data processing, pattern recognition, and visualization design. This paper provides a comprehensive overview of the current state-of-the-art on the application of AI for KV, as well as the challenges related to visualizing complex datasets. Different KV approaches, such as information visualization, ontologies and concept mapping are examined. We explore various AI techniques used to enhance these methods, including natural language processing(NLP), machine learning(ML), and deep learning(DL). Finally, we discuss the limitations and promising areas for future research in this field.

ABSTRACT. This paper presents application of artificial intelligence (AI) fuzzy logic technique for creating expert system that will be able to predict bevel angle as quality response in plasma jet cutting process. Bevel angle is significant cut quality response in plasma nonconventional manufacturing process that appears due to instability and deflection of the plasma jet during cutting and penetration in workpiece. Initial point for development of AI fuzzy logic expert system is experimental work. In this paper experimentations were conducted on aluminum sheet 5083 thickness 8 mm by varying four process parameters gas pressure, cutting speed, arc current and cutting height. Due to complexity of manufacturing process a few constraints regarding parameters values combinations in experimental plan were identified. In order to define relations between input parameters and bevel angle response fuzzy logic AI technique was applied. Prediction accuracy of developed fuzzy logic model was checked by using mean absolute percentage error (MAPE) and coeffi-cient of determination (R2) between experimental and predicted data. Also, response surface plots were created to visualize input parameters effects on the analyzed cut quality response. Presented fuzzy logic expert system enables better understanding and control of manufacturing process as well as it serves as basis for further and more detailed experimentations in this area.

ABSTRACT. Researchers have put much effort into developing effective online teaching and learning methods to support the pursuit of education during the COVID-19 pandemic. This research aims to provide insight into experiences gained from the bichronous online learning course Teaching the Teachers, intended to improve teachers' knowledge and skills in using information and communication technology. A group of 44 primary and secondary school teachers from Bosnia and Herzegovina was involved in the research, conducted in May and June 2020. The online course was available through Moodle system and digital flashcard system Memory. The teachers that used the Memory outperformed those that learned on Moodle. A novel corpus-based approach to the online discussion forum proved to be an effective tool in assessing the teachers' use of topic-related terminology. We measured the correlation between formative and summative assessments.

ABSTRACT. Abstract. As artificial intelligence (AI) becomes increasingly prevalent in the legal industry, it is essential to consider its impact on legal education and law practice. This paper examines the changes in legal education in the digital society of AI. We first draw an analogy between the importance of introducing AI in law and Justinian’s Code, which led to the development of new forms of legal learning and sophisticated academic and professional legal texts. Similarly, AI allows for analyzing large amounts of legal data and provides new knowledge about legal trends and risks. AI tools like machine learning algorithms can automate routine legal tasks and accelerate legal research. However, AI also presents challenges related to transparency, accountability, ethics, and privacy, which require new legal education and practice approaches. Law students need to be aware of the impact of AI on the law and society and be equipped with the necessary skills to work with it. Furthermore, legal professionals need to understand the implications of AI to create appropriate legal frameworks to mitigate potential risks while avoiding impeding innovation. Overall, this paper highlights the need for a new paradigm in legal education and practice that considers the evolving role of AI in the legal industry.

ABSTRACT. The first step in eliminating academic dishonesty (in e-learning systems) is to detect fraudulent activities. There are various approaches that deal with this problem, but only few of them are based on human-computer interaction (HCI). Accordingly, we have developed a novel data acquisition model that collects information about student HCI activities. This model was applied within an open-source module, named Student Activity Tracker, available as chromium web browser extension. During the COVID-19 pandemic (in academic year of 2020/2021.), an experiment was conducted with 54 volunteer participants that performed online tests from home. As a result, 500k raw logs were collected and later processed and used to develop four machine learning models. The main contribution of this research is the proof of an obvious correlation between HCI activities and fraudulent behaviour even on such a small sample. Also, we believe that this module can help other researchers to create new data sets and build new more advanced models.

ABSTRACT. Everyday activities enable a child to develop and adopt pre-reading skills which are the base for learning how to read. Digital technology and many available digi-tal tools provide different possibilities for learning, researching and acquiring knowledge. Considering how important pre-reading skills are for children and their education from the very early age, this paper researches the application of digital tools with the purpose of acquiring pre-reading skills and their impact on children in daily work in the institution of early and preschool education. This paper focuses on pre-reading skills of sound synthesis, sound analysis, syllables recognition and segmentation, creating the connection between a sound and a let-ter, connecting the first sound in a word and recognizing rhyme and rhyming. With the help of digital apps and games, children developed their pre-reading skills during the period of four months, and they improved in all tested catego-ries. The results were compared to the results of the children who did not use dig-ital tools, and it was noticed that children can adopt and develop pre-reading skills with the help of digital technology.

ABSTRACT. Knowledge tracing is critical in educational data mining to accurately predict student performance and knowledge in future interactions. In this study, we utilized the EdNet dataset to implement different classifiers and evaluate their performance in knowledge tracing. We found that the LightGBM classifier outperformed other classifiers, achieving the highest accuracy of 0.723 on the test set. Furthermore, we applied three ensemble methods, including boosting, bagging, and stacking, to investigate if they could further improve the performance of the LightGBM classifier. Our experimental results showed that the stacking method provided the most significant improvement, while the bagging and boosting methods did not yield significant improvements. These findings suggest that the LightGBM classifier, combined with the boosting method, can be an effective approach for knowledge tracing and have potential practical applications in education.

ABSTRACT. The aim of this research is to propose a fuzzy decision-making model for select-ing the most suitable text summarization method in e-learning systems, using the subjective judgments of decision makers. The criteria for selecting the optimal text summarization method in this study are based on the common characteristics of existing text summarization methods. Since there is more than one criterion for evaluating text summarization methods, it is necessary to use multi-criteria deci-sion-making methods (MCDM). The proposed approach is based on the FAHP (fuzzy analytic hierarchy process) and TOPSIS (Technique for Order of Prefer-ence by Similarity to Ideal Solution) methods. In our method, FAHP is first used to determine the weights of each criterion. Triangular fuzzy numbers are used in the FAHP method to determine the preference of one criterion over another. Then, the TOPSIS method is used to determine the final ranking of text summari-zation methods. The best method would be the one that is farthest from the nega-tive ideal solution and closest to the positive ideal solution according to the TOPSIS method. The integration of FAHP and TOPSIS methods enables effi-cient selection of the most appropriate text summarization method according to one's needs.

ABSTRACT. The basic concept of Internet of Things is the ability to upgrade everyday objects with identification, sensor, network, and processing capabilities that will enable them to communicate with each other, as well as with other devices and services via the Internet. Improved in this way, these objects become smart objects, because they require minimal or even no human intervention to generate, exchange, collect, analyze, and manage data. The numerous possibilities of IoT provide great potential for its application in various areas of human life. In this paper, a comprehensive review of IoT implementation research is given, with a special emphasis on its application in education. The opportunities and advantages that IoT provides to educational institutions and all stakeholders in the learning and teaching process are listed and categorized. Challenges that institutions face when introducing IoT in their daily work are also listed, as well as suggestions for possible solutions for each of the challenges.

ABSTRACT. Robotics has proven to be a powerful learning tool, not only as a subject of study and an engineering branch, but also for general aspects of studying STEM disciplines. There are numerous areas of engineering technique and sci-ence that are included in the STEM concepts of education based on the application of robotics, such as: sensors and actuators, voice recognition, image pro-cessing, new technologies in general, the Internet of Things, smart devices, the digital world as a whole, Industry 4.0 /5.0, etc. The inclusion of virtual programming environments and simulators in the teaching of robotics can eliminate many limitations, and enable a larger number of students to simultaneously have access to quality teaching at the same level. This paper presents an overview of several robotics programming environments and simulators, which can be an excellent tool for teaching STEM disciplines through programming, modeling and simulating different robots.

ABSTRACT. Part-of-speech tagging is one of the fundamental tasks in the field of natural language processing. It does not, in itself, solve a particular problem of nat-ural language processing, but it is a prerequisite for many other processes, such as sentiment analysis, text translation, grammar checking or speech recognition. In order to determine the efficiency of using different neural network architectures for this task, RNN and LSTM networks were imple-mented. The aim of this paper is to explore the area of part-of-speech tag-ging for the Croatian language and to show the application of neural net-works for the part-of-speech tagger of the Croatian language.

ABSTRACT. This paper presents a framework for initiating a fire inci- dents database using web scraping and impact assessment techniques. The methodology used for correct space, time, and impact association is described. The dataset used in this study was collected from media re- ports published daily by the Croatian Fire Association during a specific period. Each incident in the collected data was assigned an estimated time of occurrence and evaluated for its impact level. The study includes a comparison of the obtained results with the official reports of fire im- pact. In this context, impact refers to the extent of damage caused by the fire incident.