Hibridni događaj

Program događaja

petak, 24.5.2024 9:00 - 13:30, Camelia 2, Grand hotel Adriatic, Opatija 9:00 - 14:00 Radovi  1. I. Črnjak, M. Sokele, S. Morić (Tehničko veleučilište u Zagrebu, Zagreb, Croatia) Android Application for Measuring 5G Mobile Network Radio Parameters  At Zagreb University of Applied Sciences, as a part of the internal project ICT Green Routes for Recreation in Zagreb, a new enhanced mobile application has been developed for measuring mobile network radio parameters. Initially developed to support recreational walks and enhance student education, the application now features an enriched user experience. The paper presents the design and development of the improved user interface and added capability to measure 5G mobile network radio parameters. The application measures the radio parameters of the mobile network while recording the corresponding geographic coordinates and current measurement timestamps. Furthermore, this application operates under an open-source framework, emphasizing a commitment to transparency, collaboration, and community-driven development. The measurement of the radio parameters of the network is carried out through indicators essential for high-quality data transmission between the base station and the users (signal strength, signal quality, signal-to-noise ratio, etc.). To test the functionality of the improved application, an analysis was carried out related to the coverage of 5G mobile network on a chosen green route for pedestrian recreation in the city of Velika Gorica. 2. M. Pristavnik Vrešnjak (University of Ljubljana, Ljubljana, Slovenia), A. Perušić (University of Rijeka, Rijeka, Croatia), Ž. Emeršič, P. Peer, B. Batagelj (University of Ljubljana, Ljubljana, Slovenia) Preliminary Study on Detection of Breasts  In the realm of digital image manipulation, deep fakes, predominantly sourced from pornographic materials, present a significant challenge, especially prevalent in the form of face and body swapping techniques. This emerging issue involves substituting the faces or bodies of individuals in explicit content, using advanced methods like those demonstrated in the DeepNude application. In response, we present an approach which solves the first part of the pipeline – detection of breasts. There has been limited research regarding this biometric modality, with notable exceptions such as breast cancer identification. Due to the lack of research and the absence of open, freely available data, we developed our own dataset. Images with annotations were acquired from pornhub.com and curated by experts. Annotations include name, cup size, possible breast augmentations, ethnicity, among others. To demonstrate that dataset is challenging enough for future research we used images to train in class-agnostic way three CNN-based detection models. The results show the feasibility of not only the proposed detection approaches for the task, but also the dataset and hopefully pave the way for future applications such as supporting court decisions, enhancing virtual clothing fitting techniques, and more. 3. A. Šarčević, A. Merlin, M. Horvat (Faculty of electrical engineering and computing University of Zagreb, Zagreb, Croatia) Enhancing Programming Education with Open-Source Generative AI Chatbots  This paper describes the development of an Open-Source Generative AI Chatbot, utilizing free Large Language Models (LLM) to enhance the student learning experience for a university course in Introduction to Programming. The paper aims to provide a step-by-step guide for selecting, fine-tuning, and evaluating available models. As a first step in selecting an appropriate LLM that provides the most accurate answers while not requiring excessive computational power, the paper will include a discussion of the advantages and disadvantages of local and cloud-based available models. After selecting the best models, the next stage includes fine-tuning LLMs to answer domain-specific questions using a dataset containing essential rules, guidelines, and explanatory content in the domain of interest. The crucial aspect in selecting a model was the evaluation of answers, and in this context, both human and automatic scoring techniques are presented. Finally, it was possible to improve the performance and accuracy of the selected model by incorporating Retrieval Augmented Generation (RAG) techniques. The influence of various contributing factors, such as different vector databases, model temperatures, maximum token lengths, prompt templates, embeddings, repetition penalties, and chunking sizes, was explored. Our results show that chatbots have significant potential to improve academic support and learning efficiency, as well as personalized education in general. 4. E. Haramina, M. Paladin, Z. Petričušič, F. Posarić, A. Drobnjak, I. Botički (Fakultet elektrotehnike i računarstva, Zagreb, Croatia) Learning Algorithms Concepts in a Virtual Reality Escape Room  Although the standard way to learn algorithms is by coding, learning through games is another way to obtain knowledge while having fun. Virtual reality is a computer-generated three-dimensional environment in which the player is fully immersed by having external stimuli mostly blocked out. In the game presented in this paper, players are enhancing their algorithms skills by playing an escape room game. The goal is to complete the room within the designated time by solving puzzles. The puzzles change for every playthrough with the use of generative artificial intelligence to provide every player with a unique experience. There are multiple types of puzzles such as. time complexity, sorting algorithms, searching algorithms, and code execution. The paper presents the results of a study indicating students' preference for learning through gaming as a method of acquiring algorithms knowledge. 5. M. Jovanić, M. Čarapina (Zagreb University of Applied Sciences, Zagreb, Croatia) Application of Artificial Intelligence in the Creation of Web Content  This paper delves into the application of artificial intelligence (AI) technology in the creation of web content, with a special emphasis on A/B testing as an optimization strategy. It analyses the way technical tools like Prisma, Next.js, Tailwind, and Clickhouse contribute to the development and analysis of web applications. The importance of Large Language Models (LLM) in developing interactive interfaces and providing application performance insights is also assessed, with a focus on user behavior analysis. The paper explores how AI, specifically tools like GPT-3.5-turbo, might enhance the process of creating content for the web. The usefulness and potential of AI in generating text are looked into in the context of the continued development of digital communication strategies. Methods such as A/B testing and performance monitoring have significance in evaluating the effectiveness of AI-generated content communicating with users. The work aims to provide an understanding of how the integration of AI and present-day technologies may increase user efficiency and satisfaction in web application development. The paper demonstrates the potential and constraints of applying AI in digital communication through theoretical and empirical research, emphasizing its relevance in shaping the future of web content. 6. V. Čotić Poturić (Faculty of Informatics and Digital Technologies, Rijeka, Croatia), I. Dražić (Faculty of Engineering, Rijeka, Croatia), S. Čandrlić (Faculty of Informatics and Digital Technologies, Rijeka, Croatia) Analysis of Predictors as a Basis for the Development of an Information System for Predicting Failure in Stem Courses  As part of a doctoral thesis, an information system is to be developed to support the education system, particularly for STEM subjects. The basic functionality of the information system would be to inform the student about the risk of failure and the individualized optimal strategy to reduce this risk based on the available educational data. This article describes the analysis of correlation of predictors, i.e. independent variables, in order to identify important predictors that can be used in the design of the described information system. The article uses the basic methods of descriptive statistics in the form of calculating indicators of central tendency and dispersion, Pearson's and Spearman's correlation coefficients and Cramer's association coefficient. The main objective of this study is to create a set of relevant variables that form the basis for the construction of the information system described. 7. L. Lasic (RIT Croatia, Zagreb, Croatia), A. Radovan (Algebra University College, Zagreb, Croatia), B. Mihaljević (RIT Croatia, Zagreb, Croatia) Optimizing Machine Learning Training: A Comparative Study of Storage Types for Efficient Large Dataset Processing   This research paper explores various storage types suitable for handling large datasets within machine learning projects. The study focuses on a comparison of performance metrics, specifically focusing on the read operations from storage employing textual files, relational databases, and NoSQL databases. In the realm of machine learning projects, the significance of efficiently reading extensive datasets is very important. This research seeks to describe alternative approaches to the varying sizes of datasets, recognizing that the time required to train machine learning models is related to the efficiency of data retrieval. The methodology employed for this comparative analysis encompasses the utilization of datasets of different sizes, measurement of read operation durations, iterative measurements for precision, and computation of average read operation durations. The results show that the storage type decisions need to depend on a specific size of the dataset, as well as certain characteristics, thereby optimizing the length of the training process for machine learning models. By aligning storage choices with dataset dimensions, the results of this research help to make a better decision related to choosing the right dataset storage type, thereby contributing to the overall efficiency and efficacy of machine learning systems. 8. L. Lasić, D. Beronić, B. Mihaljević, A. Radovan (RIT Croatia, Zagreb, Croatia) Assessing the Efficiency of Java Virtual Threads in Database-Driven Server Applications  Virtual Threads represent a contemporary structured concurrency model in Java Virtual Machine (JVM) seeking to increase the performance of multi-threaded Java applications by optimizing the utilization of the operating system (OS) resources. Virtual Threads were first introduced within the OpenJDK project Loom as lightweight threads based on the native implementation of thread schedulers inside the JVM that are less reliant on OS schedulers. Within the Java Development Kit (JDK), Virtual Threads were presented as a preview feature in JDK 19/20 and became fully implemented as a part of the standard JDK 21. Given the paucity of research on the efficiency of Java’s Virtual Threads, particularly in cloud-based environments, we explored the role of Virtual Threads in enhancing Java's concurrency capabilities in the realm of database-driven cloud computing, particularly server applications with thread-per-request model and various backend databases. Our research examined Virtual Threads' efficiency in comparison with Java’s traditional threads within database-driven server applications with use cases utilizing common data frameworks to access relational and non-relational databases. From our findings and preliminary results, we propose the possible utilization of Virtual Threads in modern database-driven framework-based server applications in the cloud. 9. G. Perković, A. Drobnjak, I. Botički (Faculty of electrical engineering and computing, Zagreb, Croatia) Examining Hallucinations in Large Language Models   Large language models (LLM) are trained to understand and generate human-like language. While LLMs present a cutting-edge concept and their use is becoming widespread, hallucinations sometimes occur during their operation. Hallucinations refer to instances where the model generates inaccurate or fictitious information, deviating from factual knowledge and potentially providing responses that lack a basis in model's training data. In this paper, the ways in which LLMs generate text are examined to address the question of why hallucinations occur. The paper additional explores how existing LLM models can be leveraged to reduce the likelihood of hallucination. Alongside exploring hallucinations, this paper provides insights into the algorithms used for training LLMs, offering a clear picture of the text generation process and its effective utilization. 10. I. Živković, M. Ašenbrener Katić, V. Slavuj (Sveučilište u Rijeci, Fakultet informatike i digitalnih tehnologija, Rijeka, Croatia) Usporedba odabranih računovodstvenih alata otvorenog koda  Računovodstvo je jedan od ključnih dijelova svakog poslovanja. Omogućava praćenje financijskih transakcija, izradu financijskih izvještaja i donošenje informacijskih odluka, pruža jasnu sliku financija određene organizacije te može poslužiti kao katalizator za upravljanje resursima i pomoć pri strateškom rastu. Brzi tehnološki napredak, razvoj računovodstvenih alata, uključujući softver otvorenog koda, igra važnu ulogu u optimizaciji ovih procesa. Računovodstveni alati otvorenog koda nude alternativu komercijalnim softverima i omogućavaju organizacijama da prilagode softver svojim potrebama, bez potrebe za velikim financijskim ulaganjima. Međutim, postoje prednosti i mane takvih alata. U ovome radu analizirat ćemo te usporediti neke od najpopularnijih, većinom besplatnih open source računovodstvenih alata. Za analizu su odabrani alati GnuCash, Akaunting, Frappe Books i Dolibarr ERP-CRM. Neki od kriterija usporedbe alata su moduli, općenite informacije poput licenciranja, programski jezik implementacije, postojanje verzije za hrvatski jezik i slično. Navedeni su alati analizirani koristeći primjer zamišljenog obrta te je objašnjeno kako se unose izlazni računi. Na temelju navedenoga proučeno je korisničko sučelje i njegova interaktivnost. 11. M. Mastromatteo (Management HSQE, Field & Safety Operations, Italferr S.p.A., Firenze, Italy), A. Amelio (Department InGeo, University "G. d'Annunzio" Chieti-Pescara, Pescara, Italy) A Deep Learning Approach for Predicting Air Pollutants on the Construction Site  In recent years, the problem of air pollution has become an urgent issue causing a meaningful impact on health and environment. In urban areas, one of the main sources of pollution is air pollution on construction sites. It is characterised by multiple pollutants, among which one of the most worrying harmful substances is suspended particulate (PM2.5), causing serious damage to human health and environment. Although different monitoring systems have been recently introduced for assessing the level of air pollutants on construction sites, predicting their diffusion over time has not been explored so far, which is relevant to preserve the health of workers and people surrounding the area. To overcome this limitation, we propose a new framework based on recurrent neural networks for monitoring and predicting the spread of air pollutants on construction sites, in particular PM2.5, from known environmental conditions. The framework is composed of the following steps: (i) data pre-processing, (ii) model training, (iii) model testing, and (iv) model deployment in the construction site. Results obtained on the test set prove the reliability and usability of the proposed framework for the construction sites. 12. L. Furmanek, S. Lins, M. Blume, A. Sunyaev (Karlsruhe Institute of Technology, Karlsruhe, Germany) Developing a Hybrid Deployment Model for Highly Available Manufacturing Execution Systems  Manufacturing Execution Systems (MES) are the digital core of every industrial production and require to be always operational. Migrating the MES to the cloud can harvest benefits such as higher utilization, more flexibility, and lower costs, but also increases the risk of interruptions due to connectivity issues or cloud maintenance. To ensure high availability, we propose a hybrid MES that leverages the cloud and local fog nodes as intermediate infrastructure to guarantee an always operational system. In this paper, we specify high availability for MES and review protection objectives that a hybrid MES design needs to fulfill. Subsequently, we propose a novel hybrid MES deployment model that distinguishes the management of transactional and master data to make the best use of high available fog nodes and flexible but cost-efficient cloud resources. 13. I. Polinchev, G. Yankov, K. Zlatev, A. Aristotelov, S. Hossny (Burgas Free University, Burgas, Bulgaria) Study of the Safety in Operation of Electric Vehicles and Their Accompanying Infrastructure  This report analyzes the potential risks and specific characteristics of road accidents involving electric vehicles, paying particular attention to safety aspects that differ from traditional gasoline and diesel vehicles. A test is conducted to provide a better understanding of safety issues and develop strategies to reduce the risk of road accidents involving electric and hybrid vehicles. The possibilities for structuring and operation of the charging infrastructure of hybrid and electric vehicles are considered. The problems of the safe operation of both the vehicles and the corresponding auxiliary infrastructure are also analyzed. 14. F. Cupan, L. Klaric, V. Kirincic (University of Rijeka, Faculty of Engineering, Rijeka, Croatia), F. Mitrovic (360 Mobility, Belgrade, Serbia) Decarbonization of the Transport Sector in the Green Energy Transition  This paper presents a model of decarbonization of the transport sector applied to a smaller fleet of vehicles. The impact of individual vehicles and fleets on the environment is studied through the analysis of their carbon footprint. Emphasis is placed on electric vehicles as a sustainable solution for replacing fossil fuels and on the role of investing in the modernization of the vehicle fleet through new vehicles with lower CO2 emissions, including hybrid vehicles. The paper also investigates the obstacles that make the transition to electric vehicles difficult. 15. I. Rendulić, A. Veltruski, D. Sente, S. Tvorić, B. Tomičić (Tehničko veleučilište u Zagrebu, Zagreb, Croatia) Modeliranje i određivanje karakteristika asinkronog motora korištenjem naprednih numeričkih alata  Asinkroni kavezni motori su električni rotacijski strojevi koji se često koriste u raznim industrijama zbog svoje pouzdanosti i ekonomičnosti. Zbog svoje jednostavne građe i stabilnosti u radu, omogućena mu je široka primjena za razne procese u kojima je njihov rad potreban. U sklopu rada dan je postupak projektiranja asinkronog kaveznog motora industrijske primjene korištenjem naprednih numeričkih alata. Prema definiranim nazivnim podacima i podacima o dimenzijama statora i rotora napravljen je dizajn motora u programu Simcenter Motorsolve. Kada je dizajn motora završen, analiziran je u programu Simcenter MAGNET koji se upotrebljava za simulaciju elektromagnetskih polja koji za rad koristi metodu konačnih elemenata. Uz pomoć te metode provedena je analiza, koja se odnosila na provjeru statičke radne točke, praznog hoda i kratkog spoja prethodno dizajniranog motora. Kao rezultat te provjere, za svako stanje motora su prikazani dijagrami brzine, struje i momenta u vremenu. 16. M. Vejin, M. Đoćoš (University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Serbia), S. Bučko, J. Katona (University of Novi Sad, Faculty of Technology, Novi Sad, Serbia), S. Kojić, G. Stojanović (University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Serbia) Fabrication and Characterization of a Resistor Made of Carbon Film  Evolution of edible electronics has paved the way for groundbreaking advancements in the realm of electronics, fostering innovations that merge sustainability with functionality. As the field of edible electronics continues to flourish, the integration of biocompatible materials and ingestible designs opens doors to applications in healthcare and beyond. Development of edible LRC (inductance, resistance, capacitance) components further extends this vision, and brings us closer to a more sustainable world with green and renewable electronics that are not harmful to the human body. The research focuses on the fabrication of films incorporating activated carbon aimed to innovate edible electronics. The films were prepared using a blend composed of activated carbon, propylene glycol, and zein. These films, created by combining these materials, are the basis for the production of edible electronic components. The fabrication involved the design of the resistor, its fabrication and comprehensive characterization. Through precise fabrication techniques, these films were tailored to meet the unique requirements of ingestible electronics, showcasing the potential for customizable and biodegradable electronic elements. The findings reveal the viability of edible resistors, offering a promising pathway for sustainable and biodegradable electronic applications, effectively linking technology with environmental sustainability in the electronics domain. 17. E. Starić, N. Tanković (Faculty of Informatics, Pula, Croatia) Preliminary Study on Effects of Object-Relational Mapping on the Efficiency of Monolithic and Distributed Relational Database Systems  As there is a growing need to store and process large amounts of data for the correct operation of many applications, traditional monolithic relational databases, due to their architecture, are becoming insufficient for such applications. Distributed databases are presented as a potential solution. Distributed databases store data on multiple instances, which enables greater scalability, flexibility, database availability, and potentially better performance in data processing. Furthermore, object-relational mapping (ORM) is a programming technique that facilitates the interaction between a database and a program written in an objectoriented language. It allows writing database queries in an object-oriented paradigm instead of using the SQL programming language. Although ORMs make it much easier to write programming code, they are also known to affect database performance negatively. This paper aims to determine the impact of ORM on the performance of monolithic and distributed relational databases and to compare whether it has a greater effect on monolithic or distributed databases. We will use the TPC (Transaction Processing Performance Council)-C benchmark. MySQL will be used for the monolithic database, TiDB for the distributed database, and SQLAlchemy for ORM.

Osnovni podaci:

Pavle Ergović (Croatia), Dubravko Sabolić (Croatia), Filip Srdić (Croatia)

Prijava/Kotizacija:

Popust se ne odnosi na studente doktorskog studija.

OBAVIJEST AUTORIMA: Uvjet za objavu rada je plaćanje najmanje jedne kotizacije po radu. Autorima 2 ili više radova, ukupna se kotizacija umanjuje za 10%.

Pavle Ergović
KSET
Fakultet elektrotehnike i računarstva
Unska 3
10000 Zagreb, Hrvatska

GSM: +385 99 417 3482
E-mail: pavleergovic@gmail.com

Najbolji radovi bit će nagrađeni.
Prihvaćeni radovi bit će objavljeni u zborniku radova s ISSN brojem. Radovi na engleskom jeziku prezentirani na skupu bit će poslani za uključenje u digitalnu bazu IEEE Xplore.
.............
Postoji mogućnost da se odabrani znanstveni radovi uz određenu doradu objave u sljedećim časopisima: Journal of Computing and Information Technology (CIT), MDPI Applied Science, MDPI Information Journal, Frontiers i EAI Endorsed Transaction on Scalable Information Systems.

Opatija je vodeće je ljetovalište na istočnoj strani Jadrana i jedno od najpoznatijih na Mediteranu. Ovaj grad aristokratske arhitekture i stila već više od 170 godina privlači svjetski poznate umjetnike, političare, kraljeve, znanstvenike, sportaše, ali i poslovne ljude, bankare, menadžere i sve kojima Opatija nudi svoje brojne sadržaje. 

Opatija svojim gostima nudi brojne komforne hotele, odlične restorane, zabavne sadržaje, umjetničke festivale, vrhunske koncerte ozbiljne i zabavne glazbe, uređene plaže i brojne bazene i sve što je potrebno za ugodan boravak gostiju različitih afiniteta. 

U novije doba Opatija je jedan od najpoznatijih kongresnih gradova na Mediteranu, posebno prepoznatljiva po međunarodnim ICT skupovima MIPRO koji se u njoj održavaju od 1979. godine i koji redovito okupljaju preko tisuću sudionika iz četrdesetak zemalja. Ovi skupovi Opatiju promoviraju u nezaobilazan tehnološki, poslovni, obrazovni i znanstveni centar jugoistočne Europe i Europske unije općenito.

Detaljnije informacije se mogu potražiti na www.opatija.hr i www.visitopatija.com.