University of Muenster, Germany
Sergei Gorlatch is Full Professor of Computer Science at the University of Muenster (Germany) since 2003. Earlier he was Associate Professor at the Technical University of Berlin, Assistant Professor at the University of Passau, and Humboldt Research Fellow at the Technical University of Munich, all in Germany. Prof. Gorlatch has more than 200 peer-reviewed publications in renowned international books, journals and conferences. He obtained several Best Paper Awards at top conferences, and has been principal investigator in several international research and development projects in the field of software for parallel, distributed, Grid and Cloud systems, machine learning, and networking, funded by the European Community and by German national bodies.
Future networked applications pose very high requirements on performance, real-time behavior, and correctness. They must connect a potentially very high number of participants who interact with the application and with each other in real time, i.e., response to actions must happen virtually immediately. Challenging examples are multiplayer online computer games, advanced simulation-based e-learning, and real-time data mining. All these applications are characterized also by strict Quality of Experience (QoE) requirements, including short response times to user inputs, frequent state updates, large and frequently changing numbers of users in a single application instance. This talk will specifically address the application development process by developing a high-level design approach for challenging networked applications using modern technologies of Cloud computing and Software-Defined Networking (SDN).
aivancity School for Technology, Business & Society, Paris, France
Antoun Yaacoub is an Associate Professor and Director of the MSc in Generative Artificial Intelligence at aivancity School for Technology, Business & Society in Paris, France. He holds a Ph.D. in Artificial Intelligence and his research sits at the intersection of AI, cognitive science, and pedagogy, focusing on generative AI and large language models for education: how to generate learning content that is not just fluent but cognitively sound, linguistically well-calibrated, and explainable. He has published more than twenty papers in international conferences and journals including IJCNN, RCIS, ICCSIT, ICMV, and KES, spanning topics from AI-driven educational assessment and cognitive-framework alignment to federated learning and multi-agent systems. He is the creator of OneClickQuiz, a generative-AI-powered Moodle plugin for instant, adaptive quiz generation, adopted as both a teaching tool and a research testbed. His current work focuses on building certification and explainability methods for AI-generated educational content, aiming to make generative AI in education not just powerful, but verifiably trustworthy.
Generative AI can now write test questions, essays, and full assessments in seconds, and universities and ed-tech platforms are racing to adopt it. But speed is not the same as trust: can we actually rely on what these systems produce? Most AI-generated assessment tools optimize for producing content quickly, with little verification of whether a question tests the intended depth of thinking, whether its language matches its intended difficulty, or why it was judged "good enough" in the first place. This talk presents a line of research addressing exactly that gap. First, we align AI-generated questions to established cognitive frameworks — Bloom's Taxonomy and the SOLO taxonomy — to verify that a question actually exercises the reasoning skill it claims to target, rather than just resembling a plausible exam question. Second, we analyse the linguistic properties of AI-generated questions and feedback — readability, lexical complexity, and challenge level — to catch mismatches between intended and actual difficulty that cognitive alignment alone can miss. Third, and most recently, we move from simply scoring AI-generated content to certifying it: building explainability methods that justify, in human-interpretable terms, why a given question or piece of feedback meets a quality standard, rather than returning an opaque pass/fail label. These ideas are grounded in a real, deployed tool: OneClickQuiz, a Moodle plugin that generates adaptive quizzes instantly using generative AI, used as a live testbed for these methods in actual classrooms. The argument I want to leave the audience with is broader than education: as generative AI is embedded into higher-stakes decisions — grading, certification, content moderation, compliance reporting — the same rigor engineers demand of a network protocol (verifiable, explainable, auditable behaviour) needs to be demanded of generative AI output, not just impressive results. Educational assessment is simply an unusually clear, high-stakes proving ground for that broader problem.
More speakers to be announced.
Copyright © All Rights Reserved, The 10th International Conference on Communications and Future Internet