1. A brief review on benchmarking for large language models evaluation in healthcareLeona Cilar Budler, Hongyu Chen, Aokun Chen, Maxim Topaz, Wilson Tam, Jiang Bian, Gregor Štiglic, 2025, pregledni znanstveni članek Opis: This paper reviews benchmarking methods for evaluating large language models (LLMs) in healthcare settings. It highlights the importance of rigorous benchmarking to ensure LLMs' safety, accuracy, and effectiveness in clinical applications. The review also discusses the challenges of developing standardized benchmarks and metrics tailored to healthcare-specific tasks such as medical text generation, disease diagnosis, and patient management. Ethical considerations, including privacy, data security, and bias, are also addressed, underscoring the need for multidisciplinary collaboration to establish robust benchmarking frameworks that facilitate LLMs' reliable and ethical use in healthcare. Evaluation of LLMs remains challenging due to the lack of standardized healthcare-specific benchmarks and comprehensive datasets. Key concerns include patient safety, data privacy, model bias, and better explainability, all of which impact the overall trustworthiness of LLMs in clinical settings. Ključne besede: artificial intelligence, benchmarking, chatbots, healthcare, large language models, natural language processing Objavljeno v DKUM: 12.05.2025; Ogledov: 0; Prenosov: 0
Povezava na datoteko |
2. Improving personalized meal planning with large language models: identifying and decomposing compound ingredientsLeon Kopitar, Leon Bedrač, Larissa Jane Strath, Jiang Bian, Gregor Štiglic, 2025, izvirni znanstveni članek Opis: Background/Objectives: Identifying and decomposing compound ingredients within meal plans presents meal customization and nutritional analysis challenges. It is essential for accurately identifying and replacing problematic ingredients linked to allergies or intolerances and helping nutritional evaluation. Methods: This study explored the effectiveness of three large language models (LLMs)—GPT-4o, Llama-3 (70B), and Mixtral (8x7B), in decomposing compound ingredients into basic ingredients within meal plans. GPT-4o was used to generate 15 structured meal plans, each containing compound ingredients. Each LLM then identified and decomposed these compound items into basic ingredients. The decomposed ingredients were matched to entries in a subset of the USDA FoodData Central repository using API-based search and mapping techniques. Nutritional values were retrieved and aggregated to evaluate accuracy of decomposition. Performance was assessed through manual review by nutritionists and quantified using accuracy and F1-score. Statistical significance was tested using paired t-tests or Wilcoxon signed-rank tests based on normality. Results: Results showed that large models—both Llama-3 (70B) and GPT-4o—outperformed Mixtral (8x7B), achieving average F1-scores of 0.894 (95% CI: 0.84–0.95) and 0.842 (95% CI: 0.79–0.89), respectively, compared to an F1-score of 0.690 (95% CI: 0.62–0.76) from Mixtral (8x7B). Conclusions: The open-source Llama-3 (70B) model achieved the best performance, outperforming the commercial GPT-4o model, showing its superior ability to consistently break down compound ingredients into precise quantities within meal plans and illustrating its potential to enhance meal customization and nutritional analysis. These findings underscore the potential role of advanced LLMs in precision nutrition and their application in promoting healthier dietary practices tailored to individual preferences and needs. Ključne besede: artificial intelligence, food analysis, LLM, Ilama, GPT, mixtral, ingredient identification, ingredient decomposition, personalized nutrition, meal customization, nutritional analysis, dietary planning Objavljeno v DKUM: 08.05.2025; Ogledov: 0; Prenosov: 1
Celotno besedilo (684,70 KB) Gradivo ima več datotek! Več... |
3. Entrepreneurship at the Advancement of Digitalization and Artificial Intelligence : GEM Slovenia 2024, Executive SummaryKarin Širec, Katja Crnogaj, Barbara Bradač Hojnik, Matej Rus, Polona Tominc, 2025, izvleček Ključne besede: Global Entrepreneurship Monitor, entrepreneurship, early-stage entrepreneurial activity, economic development, entrepreneurship ecosystem, entrepreneurship policy, sustainable business, digitalization, artificial intelligence Objavljeno v DKUM: 07.05.2025; Ogledov: 0; Prenosov: 0
Celotno besedilo (9,51 MB) Gradivo ima več datotek! Več... |
4. Prompt engineering for chatbotsMladen Borovič, 2025, drugo učno gradivo Opis: The learning material describes the remarkable and rapid progress in artificial intelligence conversational systems, from simple to advanced ones that now handle complex tasks and are becoming indispensable in various industries. It emphasises that the systems are quickly improving in understanding context and providing accurate answers, which is changing how technology is used. Because of this, the author introduces the concept of prompt engineering as a key skill for effective communication with these systems and obtaining the best results. The material serves as a practical guide to mastering this skill, although the author acknowledges that further development may make this knowledge unnecessary in the future. Nevertheless, the importance of clear communication remains, as conversational systems are just an extension of our need for it in the digital world. Ključne besede: artificial intelligence, conversational systems, prompt engineering Objavljeno v DKUM: 06.05.2025; Ogledov: 0; Prenosov: 12
Celotno besedilo (1,11 MB) |
5. Human-centered ai in smart farming : toward agriculture 5.0Andreas Holzinger, Iztok Fister, Iztok Fister, Peter Kaul, Senthold Asseng, 2024, izvirni znanstveni članek Opis: This paper delineates the contemporary landscape, challenges, and prospective developments in human-centred artificial intelligence (AI) within the ambit of smart farming, a pivotal element of the emergent Agriculture 5.0, supplanting Agriculture 4.0. Analogous to Industry 4.0, agriculture has witnessed a trend towards comprehensive automation, often marginalizing human involvement. However, this approach has encountered limitations in agricultural contexts for various reasons. While AI’s capacity to assume human tasks is acknowledged, the inclusion of human expertise and experiential knowledge (human-in-the-loop) often proves indispensable, corroborated by the Moravec’s Paradox: tasks simple for humans are complex for AI. Furthermore, social, ethical, and legal imperatives necessitate human oversight of AI, a stance strongly reflected in the European Union’s regulatory framework. Consequently, this paper explores the advancements in human-centred AI focusing on their application in agricultural processes. These technological strides aim to enhance crop yields, minimize labor and resource wastage, and optimize the farm-to-consumer supply chain. The potential of AI to augment human decision-making, thereby fostering a sustainable, efficient, and resilient agri-food sector, is a focal point of this discussion - motivated by the current worldwide extreme weather events. Finally, a framework for Agriculture 5.0 is presented, which balances technological prowess with the needs, capabilities, and contexts of human stakeholders. Such an approach, emphasizing accessible, intuitive AI systems that meaningfully complement human activities, is crucial for the successful realization of future Agriculture 5.0. Ključne besede: human-centered AI, smart framing, agriculture 5.0, digital transformation, artificial intelligence Objavljeno v DKUM: 23.04.2025; Ogledov: 0; Prenosov: 4
Celotno besedilo (1,10 MB) |
6. A review on building blocks of decentralized artificial intelligenceVid Keršič, Muhamed Turkanović, 2025, izvirni znanstveni članek Opis: Artificial intelligence (AI) is one of the key technologies transforming our lives, while the transfer of knowledge and competencies from the academic sphere to the industry and real-world use cases are accelerating yearly. However, during that transition, several significant problems and questions need to be addressed for the field to develop ethically, such as digital privacy, ownership, and control. These are some of the reasons why the currently most popular approaches of artificial intelligence, i.e., centralized artificial intelligence (CEAI), are questionable, with other directions also being explored widely, such as decentralized artificial intelligence (DEAI), which aim to solve some of the most far-reaching problems. This paper aims to review and organize the knowledge in the field of DEAI, focusing solely on studies that fall within this category. A systematic literature review (SLR) was conducted using six scientific databases and additional gray literature to analyze and present the findings of 71 identified studies. The paper’s primary focus is identifying the building blocks of DEAI solutions and networks, tackling the DEAI analysis from a bottom-up approach. Future research directions and open problems are presented and proposed at the end. Ključne besede: artificial intelligence, blockchain, cryptography, decentralization, AI, DEAI, decentralized artificial intelligence Objavljeno v DKUM: 23.04.2025; Ogledov: 0; Prenosov: 2
Celotno besedilo (2,83 MB) |
7. Advancing sustainable mobility: artificial intelligence approaches for autonomous vehicle trajectories in roundaboutsSalvatore Leonardi, Natalia Distefano, Chiara Gruden, 2025, izvirni znanstveni članek Opis: This study develops and evaluates advanced predictive models for the trajectory planning of autonomous vehicles (AVs) in roundabouts, with the aim of significantly contributing to sustainable urban mobility. Starting from the “MRoundabout” speed model, several Artificial Intelligence (AI) and Machine Learning (ML) techniques, including Linear Regression (LR), Random Forest (RF), Support Vector Regression (SVR), Gradient Boosting Regression (GBR), and Neural Networks (NNs), were applied to accurately emulate human driving behavior and optimize AV trajectories. The results indicate that neural networks achieved the best predictive performance, with R2 values of up to 0.88 for speed prediction, 0.98 for acceleration, and 0.94 for differential distance, significantly outperforming traditional models. GBR and SVR provided moderate improvements over LR but encountered difficulties predicting acceleration and distance variables. AI-driven tools, such as ChatGPT-4, facilitated data pre-processing, model tuning, and interpretation, reducing computational time and enhancing workflow efficiency. A key contribution of this research lies in demonstrating the potential of AI-based trajectory planning to enhance AV navigation, fostering smoother, safer, and more sustainable mobility. The proposed approaches contribute to reduced energy consumption, lower emissions, and decreased traffic congestion, effectively addressing challenges related to urban sustainability. Future research will incorporate real traffic interactions to further refine the adaptability and robustness of the model. Ključne besede: sustainable mobility, autonomous vehicles, machine learning, roundabouts, artificial intelligence, ChatGPT Objavljeno v DKUM: 04.04.2025; Ogledov: 0; Prenosov: 1
Celotno besedilo (9,00 MB) Gradivo ima več datotek! Več... |
8. Statistically significant differences in AI support levels for project management between SMEs and large enterprisesPolona Tominc, Dijana Oreški, Vesna Čančer, Maja Rožman, 2024, izvirni znanstveni članek Opis: Background: This article delves into an in-depth analysis of the statistically significant differences in AI support levels for project management between SMEs and large enterprises. The research was conducted based on a comprehensive survey encompassing a sample of 473 SMEs and large Slovenian enterprises.
Methods: To validate the observed differences, statistical analysis, specifically the Mann–Whitney U test, was employed.
Results: The results confirm the presence of statistically significant differences between SMEs and large enterprises across multiple dimensions of AI support in project management. Large enterprises exhibit on average a higher level of AI adoption across all five AI utilization dimensions. Specifically, large enterprises scored significantly higher (p < 0.05) in AI adopting strategies and in adopting AI technologies for project tasks and team creation. This study’s findings also underscored the significant differences (p < 0.05) between SMEs and large enterprises in their adoption and utilization of AI technologies for project management purposes. While large enterprises scored above 4 for several dimensions, with the highest average score assessed (mean value 4.46 on 1 to 5 scale) for the usage of predictive Analytics Tools to improve the work on the project, SMEs’ average levels, on the other hand, were all below 4. SMEs in particular may lag in incorporating AI into various project activities due to several factors such as resource constraints, limited access to AI expertise, or risk aversion.
Conclusions: The results underscore the need for targeted strategies to enhance AI adoption in SMEs and leverage its benefits for successful project implementation and strengthen the company’s competitiveness. Ključne besede: enterprises, project management, leadership, artificial intelligence Objavljeno v DKUM: 04.04.2025; Ogledov: 0; Prenosov: 4
Celotno besedilo (1,10 MB) Gradivo ima več datotek! Več... |
9. Sensors and artificial intelligence methods and algorithms for human - computer intelligent interaction: a systematic mapping studyBoštjan Šumak, Saša Brdnik, Maja Pušnik, 2022, izvirni znanstveni članek Opis: To equip computers with human communication skills and to enable natural interaction
between the computer and a human, intelligent solutions are required based on artificial intelligence
(AI) methods, algorithms, and sensor technology. This study aimed at identifying and analyzing
the state-of-the-art AI methods and algorithms and sensors technology in existing human–computer
intelligent interaction (HCII) research to explore trends in HCII research, categorize existing evidence,
and identify potential directions for future research. We conduct a systematic mapping study of the
HCII body of research. Four hundred fifty-four studies published in various journals and conferences
between 2010 and 2021 were identified and analyzed. Studies in the HCII and IUI fields have
primarily been focused on intelligent recognition of emotion, gestures, and facial expressions using
sensors technology, such as the camera, EEG, Kinect, wearable sensors, eye tracker, gyroscope, and
others. Researchers most often apply deep-learning and instance-based AI methods and algorithms.
The support sector machine (SVM) is the most widely used algorithm for various kinds of recognition,
primarily an emotion, facial expression, and gesture. The convolutional neural network (CNN)
is the often-used deep-learning algorithm for emotion recognition, facial recognition, and gesture
recognition solutions. Ključne besede: human–computer intelligent interaction, intelligent user interfaces, IUI, sensors, artificial intelligence Objavljeno v DKUM: 31.03.2025; Ogledov: 0; Prenosov: 4
Celotno besedilo (8,70 MB) Gradivo ima več datotek! Več... |
10. Predicting corn moisture content in continuous drying systems using LSTM neural networksMarko Simonič, Mirko Ficko, Simon Klančnik, 2025, izvirni znanstveni članek Opis: As we move toward Agriculture 4.0, there is increasing attention and pressure on the productivity of food production and processing. Optimizing efficiency in critical food processes such as corn drying is essential for long-term storage and economic viability. By using innovative technologies such as machine learning, neural networks, and LSTM modeling, a predictive model was implemented for past data that include various drying parameters and weather conditions. As the data collection of 3826 samples was not originally intended as a dataset for predictive models, various imputation techniques were used to ensure integrity. The model was implemented on the imputed data using a multilayer neural network consisting of an LSTM layer and three dense layers. Its performance was evaluated using four objective metrics and achieved an RMSE of 0.645, an MSE of 0.416, an MAE of 0.352, and a MAPE of 2.555, demonstrating high predictive accuracy. Based on the results and visualization, it was concluded that the proposed model could be a useful tool for predicting the moisture content at the outlets of continuous drying systems. The research results contribute to the further development of sustainable continuous drying techniques and demonstrate the potential of a data-driven approach to improve process efficiency. This method focuses on reducing energy consumption, improving product quality, and increasing the economic profitability of food processing Ključne besede: drying, moisture prediction, big data, artificial intelligence, LSTM Objavljeno v DKUM: 21.03.2025; Ogledov: 0; Prenosov: 11
Celotno besedilo (2,99 MB) Gradivo ima več datotek! Več... |