-
Graph-Based Relationship and Cluster Modelling for Declarative Spatial Intelligence
Sokołowski · Ernst
Spatial intelligence is a field of study that focuses on the analysis and interpretation of spatial data to gain insights and make informed decisions. In this paper, we will focus on integration and analysis of spatial datasets, each containing shapes representing real-world objects, along with their attributes. The process is defined in two YAML files, one for detecting spatial relationships between objects, and another for defining clusters based on the detected relationships. To verify the proposed approach, a case study was conducted, which involved analysing the neighbourhood of a retail park in Krakow, Poland, to find the most similar ones.
-
LLM-Augmented Pedagogy: Enhancing Instructor Efficiency and Student Engagement via Automated Solution Summaries
Mazur · Ernst
-
What If We Restore to the Future? An Alternative Concept of Deadlock Recovery with Petri Nets
Grobelna · Karatkevich
Petri nets remain a well-established formalism for modelling flexible manufacturing systems, where resource contention and process interactions may lead to deadlock situations that halt the system operation. Usually, recovery policies rely on rollback mechanisms, forcing a system to return to an earlier legal state from which the recovered deadlock is still reachable, so a looping is possible. In this paper, we consider a backward and forward-looking deadlock recovery policy, introducing the concept of moving to an alternative state. Instead of just reversing, the method identifies a previous legal marking and its next legal marking, and augments the Petri net with dedicated recovery transitions that move the system directly toward this desirable state. The proposed approach preserves the original state space. The preliminary results indicate that restore-to-the-future recovery is a promising alternative to classical rollback-based methods.
-
LaTeX: Praktyczny Przewodnik
Szpyrka
LaTeX to system składu - zarówno oprogramowanie, jak i zestaw instrukcji - umożliwiający tworzenie dowolnego typu dokumentów o wysokiej jakości typograficznej. Sprawdza się szczególnie w pracy z treściami technicznymi i naukowymi. Dokumenty złożone w LaTeX-u cechuje determinizm, co oznacza, że niezależnie od systemu operacyjnego, pod którym odbywa się ich kompilacja, czy też drukarki użytej do drukowania tych materia\lów uzyskuje się ten sam, w pełni przewidywalny efekt.
-
Automated Curriculum Analysis Using Large Language Models and Knowledge Graphs
Gacek · Adrian
A well-structured curriculum is fundamental for providing students with a coherent and meaningful educational journey. However, distributed authorship, informal rules for writing syllabi, and the constant need for updates make curriculum development a highly challenging task. This paper introduces a novel framework that automates the analysis of existing curricula by detecting areas of inconsistency. It utilizes a Large Language Model to extract core concepts and prerequisite relationships directly from unstructured text in course syllabi. To ensure correctness and uniqueness, the extracted entities are linked to Wikidata, a collaborative and general-purpose knowledge graph. Subsequently, a curriculum knowledge graph is constructed based on the relationships between courses and educational concepts, laying the foundation for automated symbolic analysis. We demonstrate the effectiveness of our approach through experiments on the curriculum of the ‘Computer Science and Intelligent Systems program offered at AGH University of Krakow. The results are promising, as the tool provides actionable insights on how to improve the curriculum and avoid the most common mistakes.
-
A Systematic Review on the Applications of Uppaal
Grobelna · Gajewski · Karatkevich
This paper presents a systematic review on possible applications of the Uppaal tool. This tool, an integrated environment for the modeling, validation, and verification of real-time systems modeled as networks of timed automata, is currently used in various domains of science and engineering. A systematic review of the literature from the years 2022 and 2023 was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) procedure. The aim was to identify the current application areas of various versions of the Uppaal tool, including CORA, TIGA, SMC, and Stratego. A total of 188 studies were included in the review. Quantitative information on the distribution of research papers regarding access options, scientific databases, types of papers, and geographical location was obtained. This review highlights the need for further development of the Uppaal tool. In addition, it includes a brief comparison with other mainstream formal validation tools, explores the applicability of different Uppaal versions, and offers practical guidelines for version selection. Finally, key open challenges and their potential solutions are discussed to support future research and tool enhancement.
-
Augmentation of Parking Information in OpenStreetMap Using Aerial Imagery Analysis
Mazur · Ernst · Kosiba
OpenStreetMap is a free, editable map of the world, created and maintained by volunteers and available for use under an open license. Harvesting and maintenance of spatial data can be a challenging and time-consuming task, which can be automated using machine learning and existing data sources like aerial images. This paper reports on the creation of such a system – one that augments OpenStreetMap street parking orientation data.
-
Graph Preparation for Machine Learning-Based Road Parameter Estimation
Ernst · Zaworski · Sokołowski
The article presents a process for preparing data to build a machine learning model for estimating road parameters. It starts with loading data from an open data source and creating a graph. The subareas are then defined using administrative levels. For custom areas, the procedure for splitting roads is discussed when new areas are added. Relationships with land use categories are also identified, incorporating the road context alongside its attributes. Finally, the prepared data form the basis for modeling road parameters based on previous projects in a road inventory system.
-
Text Data Mining: A Case Study of Reddit
Sokołowski · Szpyrka · Kosiba
The article describes the progress of a project on text data mining performed by software engineering students. The article presents information about the source of the text data, how it was acquired and cleaned. The process of selecting classifiers to categorize posts is described, and the results of different models are compared.
-
Aggregation-Based Ensemble Classifier Versus Neural Networks Models for Recognizing Phishing Attacks
Gałka · Bazan · Bentkowska · Szwed · Mrukowicz · Drygaś · Zarȩba · Szpyrka · Suszalski · Obara
-
Telemetry System as a Source of Data for Machine Learning
Sokołowski · Ernst
This paper explores leveraging a telemetry system to construct a dataset for training a model to control an unmanned car. The central concept involves gathering data from the autonomous system to predict the accurate behavior of the vehicle. The telemetry system serves as a key tool in capturing real-time information, enabling the model to learn and replicate the desired performance of the unmanned car.
-
Estimation of Road Lighting Power Efficiency Using Graph-Controlled Spatial Data Interpretation
Ernst · Kotulski
Estimation of street lighting energy requirements is a task crucial for both investment planning and efficiency evaluation of retrofit projects. However, this task is time-consuming and infeasible when performed by hand. This paper proposes an approach based on analysis of the publicly available map data. To assure the integrity of this process and automate it, a new type of graph transformations (Spatially Triggered Graph Transformations) is defined. The result is a semantic description of each lighting situation. The descriptions, in turn, are used to estimate the power necessary to fulfil the European lighting standard requirements, using pre-computed configurations stored in a ‘big data’ structure.
-
How Spatial Data Analysis Can Make Smart Lighting Smarter
Ernst · Starczewski
The paper proposes the use of spatial data analysis procedures to automatically determine lighting situations to allow for easy estimation of the predicted power efficiency of lighting retrofit projects. The presented method has been applied to real-world data for a large city, covering over 50,000 lamps. Results show that the accuracy of the algorithm is promising but, more importantly, demonstrate that the traditional estimation methods, based on the existing infrastructure parameters, can lead to highly biased results compared to the proposed method.