- informatik-vortraege - lists.rwth-aachen.de

Einladung: Informatik-Oberseminar Herr Sascha Martin Welten
by Maaßen, Leany 06 Jan '25

06 Jan '25

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 17.01.2025, 14:00-15:00 Uhr Ort: Raum 5053.2 (großer B-IT-Hörsaal)/Informatikzentrum, Ahornstraße 55 Referent: Herr Sascha Martin Welten, M. Sc. Lehrstuhl Informatik 5 Thema: Methods for Practical Data Sharing and Decentralised Analytics: An Integrated Approach Abstract: This dissertation addresses the challenges of sharing data across organisational boundaries, which is crucial for scientific progress, particularly with the rise of AI and data-driven approaches. While data sharing is essential for innovation, concerns around privacy, security, and regulatory compliance often limit access to valuable data. This work introduces an integrated framework that combines privacy-preserving methods with Decentralised Analytics (DA), addressing key requirements such as trust, controlled access, and data heterogeneity. The research develops a DevOps-inspired lifecycle, 'DAOps', for managing shared data analysis, implemented in a novel platform, PADME (Platform for Analytics and Distributed Machine Learning for Enterprises). PADME supports various types of analysis and enables decentralised model training, offering a privacy-preserving alternative to traditional centralised methods. The platforms evaluation in healthcare and hydrology confirms its operational readiness and applicability to real-world research scenarios. Key findings highlight the importance of common data schema standards and balanced data distributions for successful data sharing. Ultimately, this dissertation contributes to the practical realisation of data-sharing concepts, promoting secure, decentralised data analysis and fostering data-driven research collaborations. Es laden ein: die Dozentinnen und Dozenten der Informatik _______________________________ Leany Maaßen RWTH Aachen University Lehrstuhl Informatik 5, LuFG Informatik 5 Prof. Dr. Stefan Decker, Prof. Gerhard Lakemeyer Ph.D., JunProf. Dr. Sandra Geisler Ahornstrasse 55 D-52074 Aachen Tel: 0241-80-21509 Fax: 0241-80-22321 E-Mail: maassen(a)dbis.rwth-aachen.de

1 0

Einladung: Informatik-Oberseminar Alexander Kugler
by Kugler, Alexander 18 Dec '24

18 Dec '24

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Dienstag, 7. Januar 2025, 16.00 Uhr Ort: Lehrstuhl i11, Im Süsterfeld 9, Aachen, Raum R259 im 2. OG Referent: Alexander Kugler M.Sc. Chair of Computer Science 11 - Embedded Software Thema: Test Case Generation from Natural Language Requirements for Embedded Systems with Semantic Role Labeling Abstract: In this talk, the application of Semantic Role Labeling (SRL) for the generation of test cases from natural language requirements for embedded systems is explored. The approach, labelled Test Generation with Semantic Role Labeling (TG-SRL), is composed of five stages and combines machine learning with a rule-based approach. Information extracted via SRL is initially aggregated into logical expressions before being translated into First-Order Logic (FOL) formulae. Test case generation is achieved using Satisfiability Modulo Theory (SMT) solving. By modifying the SMT instance according to defined tactics, a test suite is generated. The thesis concludes with an evaluation of TG-SRL using a mutant-based strength analysis, and a comparison with the Nat2Test approach from Carvalho et al. TG-SRL performs favorably and provides valuable insights into employing Natural Language Processing (NLP) methods, and in particular SRL, in the field of test case generation. The methods and concepts presented in this thesis have been implemented in a publicly available research framework. Es laden ein: die Dozentinnen und Dozenten der Informatik

1 0

[sekretariate] Einladung: Informatik-Oberseminar Kevin Batz
by Elke Ohlenforst 11 Dec '24

11 Dec '24

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 20.12.2024, 13:00 Uhr Ort: Raum 9222 (E3, Informatikzentrum, Ahornstr. 55) Referent: Kevin Batz, M.Sc. Lehrstuhl für Softwaremodellierung und Verifikation (Informatik 2) Thema: Automated Deductive Verification of Probabilistic Programs Abstract: We study both foundational and practical aspects of the automated deductive verification of discrete probabilistic programs. Our special emphasis is on the verification of possibly unbounded loops by quantitative loop invariants. We build upon Kozen's, McIver & Morgan's, and Kaminski's weakest preexpectation calculus for reasoning about expected outcomes such as the probability of terminating in some postcondition or the expected final value of a program variable. The weakest preexpectation calculus replaces predicates from classical program verification by more general expectations - functions mapping program states to numbers instead of truth values. We study a syntax for specifying expectations, providing a foundation for automated deductive probabilistic program verifiers. We prove that our syntax is expressive and hence obtain a relatively complete verification system for reasoning about expected outcomes. We then present three different approaches for automating the verification of bounds on expected outcomes of linear loops, all of which combine weakest preexpectation reasoning with Satisfiability Modulo Theories (SMT) techniques: 1. We revisit two well-established verification techniques for transition systems, k-induction and bounded model checking (BMC), in the more general setting of bounding least fixpoints of functions over complete lattices, yielding latticed k-induction and latticed BMC. Instantiating our latticed techniques with the weakest preexpectation calculus enables the fully automatic verification of linear loops taken from the literature. 2. We present a counterexample-guided inductive synthesis approach for the synthesis of quantitative loop invariants. This enables the fully automatic verification of bounds on both expected outcomes and expected runtimes. Our implementation synthesizes quantitative invariants of various linear loops taken from the literature, can beat state-of-the-art probabilistic model checkers on finite-state programs, and is competitive with modern tools dedicated to invariant synthesis or expected runtime reasoning. 3. We present PrIC3 - the first truly quantitative extension of the state-of-the-art qualitative symbolic model checking algorithm IC3 to symbolic model checking of Markov decision processes. By marrying PrIC3 with the weakest preexpectation calculus, we obtain a symbolic model checking algorithm for probabilistic programs. Alongside, we present an implementation of PrIC3 featuring IC3-specific techniques such as generalization. ------------------------------------------------------------------------ Es laden ein: die Dozentinnen und Dozenten der Informatik _______________________________________________ informatik-sekretariate Mailingliste -- informatik-sekretariate(a)lists.rwth-aachen.de Um den Newsletter abzubestellen, senden Sie bitte eine E-Mail an informatik-sekretariate-leave(a)lists.rwth-aachen.de

1 0

Einladung: Informatik-Oberseminar Patrick Schmidt
by Schmidt, Patrick 10 Dec '24

10 Dec '24

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Mittwoch, 18. Dezember 2024, 13:15 Uhr Ort: Raum 118, Erweiterungsbau 3, Ahornstr. 55 Referent: Patrick Schmidt M.Sc. Lehrstuhl für Informatik 8 Thema: Intrinsic Optimization of Maps Between Surfaces Abstract: This talk presents a series of works focusing on the construction and optimization of high-quality maps between surfaces in 3D. These surface maps establish point-to-point correspondences between multiple 3D models, enabling applications such as attribute transfer, shape interpolation, shape co-analysis, compatible remeshing, and more. We introduce a novel framework to generate surface homeomorphisms—maps that guarantee continuity and bijectivity—while minimizing intrinsic mapping distortion. Along the way, we will explore constructions in spherical and hyperbolic geometry and discuss practical solutions for continuous optimization of geometric problems using automatic differentiation. Es laden ein: die Dozentinnen und Dozenten der Informatik

2 1

Einladung: Informatik-Oberseminar Herr Alexander Tobias Neumann
by Maaßen, Leany 04 Dec '24

04 Dec '24

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 13.12.2024, 12:30-13:30 Uhr Ort: Raum 5053.2 (großer B-IT-Hörsaal)/Informatikzentrum, Ahornstraße 55 Referent: Herr Alexander Tobias Neumann, M. Sc. Lehrstuhl Informatik 5 Thema: Chatbots as Professional Companions in Large-Scale Community Information Systems: Integrating Chatbots in Educational Ecosystems Abstract: Online communities are fundamental for knowledge sharing, collaboration, and professional development in our digital age. However, challenges like information overload and low engagement hinder their effectiveness. New members often struggle to find accurate information from multiple sources. Chatbots with natural language understanding and automation capabilities can solve these challenges. By providing personalized guidance, improving efficiency, and integrating with external knowledge bases, chatbots enhance the user experience in online communities. Unfortunately, community members often lack the necessary resources or expertise to create bots that optimize their practices. This dissertation follows a design science approach that empowers communities to create and maintain customized chatbots collaboratively. We introduce a model-driven Social Bot Framework that allows community members to participate in bot creation and evaluation. This framework incorporates process mining techniques and leverages Large Language Models (LLMs) to enhance bot performance. The dissertation also presents a catalog of success factors for bots, providing a multi-dimensional model for assessing chatbot effectiveness within communities of practice. Through a co-design process involving real-world use cases, we have produced, demonstrated, and evaluated chatbots as artifacts, showcasing their impact and usefulness. Our research explores the application of chatbots in Technology Enhanced Learning (TEL). We developed a series of mentoring bots that address various aspects of the educational process. These include systems for collaborative knowledge building, providing personalized feedback on writing tasks, and facilitating quiz-based learning. We also investigated the integration of gamification elements to enhance user engagement and motivation in educational contexts. Furthermore, we leveraged LLMs to provide more sophisticated, personalized learning support. The evaluations, conducted through user studies, surveys, and performance metrics, demonstrate the effectiveness and user acceptance of our developed chatbots. Our research promotes the democratization of bot-building, enabling communities to leverage chatbot potential. This represents a significant advancement in the self-governance and development of online communities in response to changing needs and opportunities. Ultimately, this work pushes the boundaries of current community practices, opening up new horizons where bots are professional companions supporting the success of online communities. Es laden ein: die Dozentinnen und Dozenten der Informatik _______________________________ Leany Maaßen RWTH Aachen University Lehrstuhl Informatik 5, LuFG Informatik 5 Prof. Dr. Stefan Decker, Prof. Dr. Matthias Jarke, Prof. Gerhard Lakemeyer Ph.D., JunProf. Dr. Sandra Geisler Ahornstrasse 55 D-52074 Aachen Tel: 0241-80-21509 Fax: 0241-80-22321 E-Mail: maassen(a)dbis.rwth-aachen.de

1 0

Einladung Informatik-Oberseminar
by Mokhtarian, Armin 02 Dec '24

02 Dec '24

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 6. Dezember 2024, 16.00 Uhr Ort: Raum 259, Im Süsterfeld 9, 52072 Aachen Referent: Armin Mokhtarian M.Sc. Lehrstuhl Informatik 11 Thema: CPM Remote: A Remote Access to a Scaled Experimental Testbed for Connected and Automated Vehicles Abstract: Small-scale testbeds are used in research and development to provide a controlled setting for testing new concepts. They are affordable and can be duplicated, making them useful for both education and research purposes. To address the limitations of small-scale testbeds, particularly their accessibility, we introduce CPM Remote, a web-based platform that provides real-time access to the Cyber-Physical Mobility (CPM) Lab, a scaled testbed at RWTH Aachen University. This innovative approach eliminates the need for users to travel, purchase, or reconstruct physical testbeds. CPM Remote incorporates a built-in editor, a simulation environment, and a real-time interface to the CPM Lab, extending accessibility to a global audience. Users have the flexibility to conduct their own experiments or choose from two ready-to-use applications, enabling them to actively engage with the field of Connected and Automated Vehicles (CAVs). The educational application, CPM Academy, offers students an immersive learning experience by presenting a package delivery service scenario to develop algorithms within the context of CAVs. For research purposes, we introduce CPM Olympics, an annual competition that offers participants real-world scenarios to benchmark their motion planning algorithms. In our evaluation of the platform and its applications, we carried out multiple user studies and a technology acceptance analysis to pinpoint potential areas for enhancement. The feedback from each study was considered during the development of CPM Remote to enhance its feature set and usability. Based on the successfully conducted iterations of the CPM Academy and CPM Olympics, we demonstrate how CPM Remote bridges the gap between simulations and costly real-world tests, democratizing access to experimental research infrastructure.

1 0

Einladung: Informatik-Oberseminar Alexander
by DSME-Office 29 Nov '24

29 Nov '24

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Dienstag, 3. Dezember 2024, 12.00 Uhr Ort: Technologiezentrum Aachen, Seminarraum Halle F1 Referent: Alexander von Rohr M.Sc. Institut für Data Science im Maschinenbau (DSME) Thema: Probabilistic Optimization for the Control of Dynamical Systems Abstract: Controlling dynamical systems is a key aspect of automation technology. However, designing resource-efficient controllers for each system is challenging and costly. Machine learning can help automate this process, but learning-based control requires data-efficient and robust methods. We frame learning-based control as a probabilistic optimization problem, proposing novel methods for decision-making under uncertainty and learning-based control algorithms with formal guarantees. First, we address controller tuning using Bayesian optimization to efficiently improve candidate controllers, validated on synthetic problems, simulations, and hardware experiments. Second, we propose three probabilistically robust control methods for uncertain, data-driven models, addressing probabilistic stability, aleatoric uncertainty, and event-triggered learning for time-varying dynamics. Es laden ein: die Dozentinnen und Dozenten der Informatik

1 0

Einladung: Informatik-Oberseminar Herr Mohamed Behery
by Maaßen, Leany 22 Nov '24

22 Nov '24

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Donnerstag, 05.12.2024, 13:30-14:30 Der öffentliche Vortrag findet hybrid statt: Raum: Raum 5053.2 (B-IT-Hörsaal)/Informatikzentrum, Ahornstraße 55 Zoom: https://rwth.zoom-x.de/j/64360669610 Referent: Herr Mohamed Behery, M. Sc. LuFG Informatik 5 (KBSG) Thema: Digital Shadows for Industrial Robotic Agents in the World Wide Lab Abstract: The digital transformation of production has become imperative to increase efficiency, adaptability, and innovation required to face the rapidly changing market requirements, supply chain disruptions, and global competition. In addition to initiatives such as Industry 4.0 and the Industrial Internet of Things (IIoT), the Internet of Production (IoP) aims to establish a world-wide network of research and industrial partners through the World Wide Lab (WWL). Such a network requires standardized communication protocols, which need to remain flexible to accommodate the different standards, use cases, and contexts. To this end, the WWL uses Digital Shadows (DSs) to represent the different entities: parts, products, machines, processes, and companies. DSs are context and purpose-specific representations of entities in the WWL, which enables them to replace the omnipresent Digital Twins by representing different aspects of the denoted entities. The thesis offers a road map for achieving the WWL as an application on top of the IoP. The thesis focuses on industrial robotics processes, which involve teleoperated robots or human-robot collaborative assembly and are not fully integrated in the WWL as they lack a standard DS for task representation. To bridge this gap, the thesis proposes using Behavior Trees (BTs), which offer a flexible, modular, and composable task representation for industrial robotics processes. BTs emerged in the video game industry and were adopted by the roboticists but do not yet fulfil all the requirements to be a standard DS. The presented work offers a plan to fulfill these requirements. The thesis introduces a framework for extracting and creating DSs for robot teleoperation tasks in industrial processes. It starts by collecting data to extract discrete actions. After that, it learns a library of BTs that represent the commonly executed behaviors. This allows a user to compose more complex behaviors using the learned BTs. The different modules of the framework are evaluated qualitatively and quantitatively. The thesis also proposes and demonstrates three BT extensions increasing BT flexibility to accommodate more use cases. They allow a BT to handle highly dynamic production environments, where it interacts with heterogeneous agents such as humans or robots with different capabilities or knowledge. The dissertation concludes with a discussion that evaluates the applicability of the proposed framework and extensions to different industrial scenarios giving recommendations for future work. Es laden ein: die Dozentinnen und Dozenten der Informatik _______________________________ Leany Maaßen RWTH Aachen University Lehrstuhl Informatik 5, LuFG Informatik 5 Prof. Dr. Stefan Decker, Prof. Dr. Matthias Jarke, Prof. Gerhard Lakemeyer Ph.D., JunProf. Dr. Sandra Geisler Ahornstrasse 55 D-52074 Aachen Tel: 0241-80-21509 Fax: 0241-80-22321 E-Mail: maassen(a)dbis.rwth-aachen.de

1 0

Einladung Informatik-Oberseminar Dr. Till Hofmann, 27. November 2024,
by Stephanie Jansen 21 Nov '24

21 Nov '24

********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * *********************************************************************** Zeit: Mittwoch, 27. November 2024, 13:00 Uhr Ort: Raum 9222, E3, Informatikzentrum Referent: Dr. Till Hofmann; Lehrstuhl Informatik 6 Thema: Uncertainty in Robotics Abstract: Robotic systems must navigate a world full of unpredictability, facing challenges from dynamic environments, noisy sensors, and imperfect actuators. A robot’s environment is constantly changing, influenced by interactions with other robots and humans. Sensors, with their inherent limitations and noise, provide only an imperfect view of the world. Finally, actuators—susceptible to noise, wear, and mechanical failure—introduce further uncertainty, complicating the robot’s ability to execute precise actions. In this lecture, I will discuss how these challenges can be addressed through probabilistic robotics, where uncertainty is explicitly represented using probability distributions to describe the robot's estimated state of the world. I will first explore how fundamental robotics tasks—such as localization, mapping, and control—are inherently uncertain and demonstrate the benefits of probabilistic methods. In the second part, I will introduce Bayes filters, a general algorithm for belief computation that serves as the theoretical foundation for many robotics algorithms. Es laden ein: die Dozentinnen und Dozenten der Informatik -- Stephanie Jansen Faculty of Mathematics, Computer Science and Natural Sciences Chair of Computer Science 6 ML - Machine Learning and Reasoning RWTH Aachen University Theaterstraße 35-39 D-52062 Aachen Tel: +49 241 80-21601 sek(a)ml.rwth-aachen.de www.hltpr.rwth-aachen.de

1 0

Einladung: Informatik-Kolloquium Nitesh Goyal (Google)
by Martin Henze 20 Nov '24

20 Nov '24

+********************************************************************** * * * Einladung * * * * Informatik-Kolloquium * * * +********************************************************************** Zeit: Mittwoch, 27. November 2024, 10.30 Uhr Ort: AH V, Ahornstr. 55 Referent: Nitesh Goyal, PhD Google, New York Thema: Leveraging AI Responsibly in Sensemaking for Successful Human-AI Workflows Abstract: My research vision is to enable expert and non-experts to successfully make sense of complex world problems. As a Human-Computer Interaction researcher, I iteratively focus on studying how sensemaking is performed to identify challenges in collaborative data analytics, design tools using computational techniques that overcome these challenges and evaluate my designs using human participants to inform subsequent designs. In particular, I have focussed on three domains: criminal justice, ML Development and Online Harassment. Solving crimes correctly is a critical and life-altering problem where intelligence analysts are constantly struggling against their biases. ML engineering and development requires constantly struggling with fairness and equity issues related to datasets and algorithmic outcomes. Managing online harassment requires emotional and physical labor by targets of harassment while making sense of who is harassing them and why? There are many similarities across these 3 domains and despite, recurring themes of how AI should be designed responsibly to support these use cases/users, we have barely started to scratch the surface. In this talk, I introduce the notion of Sensemaking Translucence into biases, fairness and equity related challenges. I then provide examples of how AI can support Sensemaking Translucence. My work finally makes the case that it is important to design from a human centered perspective by leveraging AI to support these Human AI Collaboration workflows. Es laden ein: die Dozentinnen und Dozenten der Informatik -- Prof. Dr. Martin Henze Security and Privacy in Industrial Cooperation (SPICe) RWTH Aachen University Im Süsterfeld 9 52072 Aachen, Germany henze(a)spice.rwth-aachen.de https://www.spice.rwth-aachen.de/

1 0