informatik-vortraege September 2024

informatik-vortraege@lists.rwth-aachen.de

4 Teilnehmer
4 Diskussionen

[Informatik-assistenten] Einladung Informatik-Oberseminar Lukas Netz

von Netz, Lukas

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Montag, 07.10.2024, 15:00 Uhr Ort: Informatikzentrum der RWTH Aachen University, Gebäude E3, 2. Etage, Raum 9222 Referent: Lukas Netz, M.Sc. Lehrstuhl für Software Engineering Thema: Model-Driven Method for the Development of Full-Size Web-Based Information Systems Abstract: The implementation of information systems poses a variety of challenges for developers and requires a large number of experts from different application domains. This work presents a methodology in the area of generative model-driven development (MDD) for web-based information systems. The methodology consists of three basic building blocks: LLM4CD, CD2GUI and MontiGem. The first building block LLM4CD is based on generative artificial intelligence (LLMs) and transforms natural language requirements and domain descriptions into class diagrams in CD4A notation. The second building block CD2GUI uses these class diagrams and derives models for user interfaces in GUI-DSL notation. The last building block is MontiGem. It uses both types of models to generate a data-centric information system. An important part of this work is the adaptability of the generated artifacts, this mainly concerns the possibility for modelers to adapt the models generated by the AI and CD2GUI, and to adapt and extend the code generated by MontiGem without having to edit the generated code or models directly. This makes iterative work with the methodology possible without discarding adapted code or adapted models. It is precisely this adaptability that facilitates the transition from an underspecified modeled software prototype to a full-size model-driven information system that can be actively used in industry. In addition to the three building blocks, this thesis also presents the domain-specific modeling languages GUI-DSL and CD4A. The methodology is evaluated and discussed both in parts and as a whole. The practical applicability of the methodology is demonstrated using the case study MaCoCo, an information system used by RWTH Aachen University, which serves as a robust benchmark for the validation of the methodology. Es laden ein: die Dozentinnen und Dozenten der Informatik _______________________________________________ Informatik-assistenten mailing list -- informatik-assistenten(a)lists.rwth-aachen.de<mailto:informatik-assistenten@lists.rwth-aachen.de> To unsubscribe send an email to informatik-assistenten-leave(a)lists.rwth-aachen.de<mailto:informatik-assistenten-leave@lists.rwth-aachen.de> https://lists.rwth-aachen.de/postorius/lists/informatik-assistenten.lists.r…

1 Woche, 2 Tage

Einladung: Informatik-Oberseminar Jan Böker

von Jan Böker

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 27. September 2024, 10:00 Uhr Ort: Raum 9222 (Informatikzentrum E3, Ahornstraße 55) Referent: Jan Wilhelm Böker, M.Sc. Lehrstuhl für Logik und Theorie diskreter Systeme (Informatik 7) Thema: Structural Similarity of Graphs: Graphons and Homomorphism Densities Abstract: The abundance of graph-structured data in domains like bio-informatics has led to the popularity of both graph embeddings and graph neural networks as techniques for machine learning on graphs. Advancements in this field, however, are mostly driven by intuition and empiricism, and there is a lack of a theoretical foundation that explains both the success and the limits of these techniques. The use of tree homomorphism counts is appealing in order to define graph embeddings that not only are well-founded theoretically but also efficiently computable. Graphons emerged as the limit objects in the theory of dense graph limits and are a generalization of graphs to measurable functions. We develop a framework for graphon similarity based on iterated degree measures, which were recently introduced by Grebík and Rocha in a generalization of Color Refinement, a popular heuristic algorithm for graph isomorphism testing, to graphons. We show that, besides tree homomorphism densities, also message-passing graph neural networks integrate into this framework. We integrate a metric variant of Color Refinement into our framework. Furthermore, we introduce the tree cut distance, a variant of the cut distance from the theory of dense graph limits. Our framework allows us to prove that these four concepts all yield the same notion of similarity of graphons. In the special case of graphs, our framework gives us four similarity measures that are not only equivalent but can also be computed efficiently. We also briefly study the computational complexity of the homomorphism reconstructability problem, which asks whether a given vector of natural numbers can actually be realized as homomorphism counts to a graph. We show that this problem is NP-hard even in a severely restricted form. The k-dimensional Weisfeiler-Leman algorithm is a generalization of Color Refinement to higher dimensions. To initiate the development of a higher-dimensional variant of our framework for graphon similarity, we generalize this algorithm and various of its characterizations to graphons and prove their equivalence. Es laden ein: die Dozentinnen und Dozenten der Informatik

2 Wochen, 4 Tage

Einladung: Informatik-Oberseminar Johannes Theissen-Lipp geb. Lipp

von Theissen-Lipp, Johannes

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Mittwoch, 25. September 2024, 11.00 Uhr Ort: COMSYS E3/ Raum 9222, Ahornstraße 55, 52074 Aachen (2359/222, 2.OG) Referent: Johannes Theissen-Lipp M.Sc. Computer Science 5 – Information Systems and Databases Thema: Semantic Foundations of Dataspaces Abstract: Digital transformation is rapidly reshaping industries, organizations, and societies around the world, driven by the exponential growth of data from the Internet and the Web. This data, encompassing both open and closed resources, drives innovation and provides competitive advantages. However, there is a growing demand for data sovereignty, where individuals and organizations control their data throughout its lifecycle. Initiatives such as dataspaces, funded with €4-6 billion in Europe, and the Internet of Production focus on data autonomy and integration, but interoperability and common understanding remain critical challenges. The fields of Semantic Web technologies and FAIR data provide frameworks for addressing these challenges. These semantic technologies enable seamless data exchange and meaningful interpretation to maximize the potential of data, yet they are not sufficiently utilized in dataspaces. This thesis explores the semantic foundations of dataspaces, aiming to integrate semantic technologies to improve data interoperability and facilitate common understanding among stakeholders. It also advances semantic technologies to enhance their impact. This work proposes new concepts, principles, best practices, and solutions to improve data management in dataspaces. The key contributions of this thesis include: - A Tailored Data Lifecycle Model for Dataspaces: This model addresses the unique challenges and requirements of dataspaces, ensuring a streamlined and structured approach to managing data and metadata throughout their lifecycle. - Principles of Information Models: These models define a common core in dataspaces, providing a structured representation of data, services, participants, and interactions, and serving as a foundation for data sovereignty and interoperability mechanisms. - Extended Data Access Principles: These principles cover various types of heterogeneity, improving data accessibility by managing resource constraints, handling dynamic data changes, and providing rich metadata to increase the quantity and quality of information in dataspaces. - Methods for Improving Common Domain Understanding: This includes techniques for identification, best practices for vocabulary development, involvement of non-experts, effective recommendation systems, evolution of semantic information over time, and improved validation techniques to enhance data interoperability and (re)use in dataspaces. By combining these contributions, this thesis confidently integrates semantic technologies into dataspaces. This integration enhances data and service interoperability and facilitates common understanding within these environments, promoting proper data management and reuse throughout the data lifecycle. Our work provides valuable insights into the effective management of information within the digital ecosystem of dataspaces, contributing to the advancement of knowledge in today's data-driven digital transformation. Es laden ein: die Dozentinnen und Dozenten der Informatik

3 Wochen, 1 Tag

Einladung: Informatik-Oberseminar Marco Lutz geb. Grochowski

von Grochowski, Marco

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Dienstag, 10. September 2024, 10.00 Uhr Ort: Im Süsterfeld 9, 52072 Aachen, Raum 259 Referent: Marco Lutz geb. Grochowski M.Sc. Informatik 11 - Embedded Software Thema: Test Suite Generation and Augmentation for Reconfigurable Industrial Control Software in the Internet of Production Abstract: With the advent of Industry 4.0 and the digitally networked factory, cyber-physical production systems (CPPSs) are reconfigured frequently along their life cycle to adapt to changing customer requirements or market demands. Such reconfigurations are not limited to the hardware but also affect the software of the programmable logic controllers (PLCs) driving these plants. While verification and testing are two techniques capable of alleviating the risk of introducing errors in production code, it is no longer sufficient to rely only on the results obtained by these methods during the commissioning of the CPPS. Even minor incremental reconfigurations to the PLC’s software during the operational phase of the life cycle may introduce regressions that can be quickly overlooked by a developer and therefore need to be reverified. The goal of this thesis is to provide a “push button” analysis for generating test cases after a static reconfiguration. The generated test cases can be injected and monitored during maintenance or virtual commissioning to observe the impact of reconfiguration on the CPPS by the developer. In order to reduce redundancy in test suite generation (TSG) after a structural reconfiguration to the PLC software, symbolic summaries of specific parts of the program should be cached and reused to benefit subsequent analysis. While automatic TSG is an established technique used to generate test suites adhering to structural coverage metrics of PLC software, the generated test suite might not anymore be adequate enough with regards to the coverage metric to ensure the absence of regressions. An indispensable part of regression testing (RT) is test suite augmentation (TSA), which guides the TSG toward the reconfigured behavior and increases the chances of deriving difference-revealing test cases which expose behavioral differences between the program and its reconfigured version. The derivation of new test cases is required to uncover potential regressions after a reconfiguration. To this end, the contributions of this thesis include * heuristics for the scalability of the existing TSG for PLC software, * the reuse of symbolic summaries during TSG of reconfigured PLC software, * and the concept of executing the old and new version of a reconfigured PLC software in one unified program version during TSA. These contributions are evaluated on selected domain-specific benchmarks of varying difficulty, such as the PLCopen Safety suite and the Pick and Place Unit (PPU). Es laden ein: die Dozentinnen und Dozenten der Informatik

1 Monat

2024

2023

2022

2021

2020

2019

2018

informatik-vortraege September 2024