KeynotesSTAF 2025

Recent advances in applied category theory and the maturation of proof assistants such as Rocq [1] (formerly Coq), Isabelle/HOL, and Lean have opened new frontiers for formalizing complex mathematical structures. Among these, the algebraic approach to graph rewriting—rooted in rich categorical foundations—offers a compelling case for mechanized reasoning. This presentation focuses on recent advancements in creating an interactive framework based on Rocq for certifying diagrammatic proofs in graph transformation theory. At the core of this work is the Instantiation Calculus, a new formalism developed as part of the ANR CoREACT project [2]. This calculus utilizes a specific type of hyperdoctrines to rigorously encode diagrammatic proofs. We will demonstrate how this calculus facilitates the formalization of essential reasoning patterns in theoretical graph transformation, while also exploring its potential for broader applications in graph rewriting theory.

Modern cyber-physical systems (CPS) such as cars, production plants, or energy networks integrate computational processes with physical processes. Due to their size and complexity, they are modelled by engineers with multiple, heterogeneous tools and modelling languages, which provide role-specific views. At the same time, these systems can only function if all parts interact seamlessly, which is only possible if models are consistent with each other. Consistency management becomes the key task in managing the complexity of modern, cyber-physical system. In this talk, I will present ongoing research in the “Convide” project (https://www.sfb1608.kit.edu/) and the Vitruvius approach (https://vitruv.tools/), which demonstrate that model-driven technologies and especially transformations are not a thing of the past, but at the core of solutions to improve the adaptability, reliability, and security of modern CPS.

TBA

Public administrations are under pressure to modernize outdated, document-heavy processes while coping with a shortage of skilled staff and rising citizen expectations. Generative AI (GenAI) presents new opportunities to automate case handling and accelerate digital transformation. This talk presents a technical and process-oriented perspective on how Agentic AI can move beyond document understanding to autonomously support and execute public sector workflows.

Consistency management in modeling: a broad topic. Individual notions of this are rarely the same but depend on the needs of the respective context. What they have in common is that it is about consistency within one or between multiple models. But what does it mean for models to be consistent? And do we only want to document it, validate it, or even enforce it automatically? The everyday experience of many engineers is still that notions of consistency tend to be implicit or at least unformalized. Checking or maintaining it is often a manual task. In this talk, we will combine views on the topic of consistency management from research and industrial practice: What kinds of consistency and consistency management are of actual interest? How do theoretical concepts relate to approaches applied in practice and which of these approaches work well despite theoretical limitations? And where do we need better concepts and research to address requirements from practice? We will bring together a research-oriented view of integrated and distributed consistency validation and preservation with practical experience from the PREEvision tool. It is an integrated tool for holistic development of electric/electronic architectures that covers the information needs of many engineers in a single model. However, even such an integrated approach requires internal mechanisms for ensuring consistency of the complex information inside that single model. The consistency management concepts and techniques it implements will give us an impression of how foundational concepts are combined with sufficient pragmatics and what practical benefits have already been achieved. Beyond this state of practice, we will take a look at increasing demands on consistency management in modularized, distributed development, deployment, and update. Finally, based on the resulting challenges, we will highlight where foundational research is needed to ensure the future success of consistency management in practice.

This session kicks off the Agile MDE track with a keynote on the evolving role of Agile Model-Driven Engineering (MDE) in adaptive systems. It explores cutting-edge applications, from digital signal processing to integrating large language models (LLMs) in agile software development workflows. Presentations focus on the synergy between agility and automation in modern engineering environments.

Large language models (LLMs), such as Copilot and ChatGPT, have received considerable attention from software engineering (SE) researchers and practitioners. Recent SE publications show that applying LLMs to specific SE tasks is a major trend. As such, LLMs have the potential to transform software development and evolution processes. In this talk, we look at how LLMs are being applied to selected SE tasks and discuss the implications of this trend for model-driven software development.

Till now development of a new software language is far too often an artistic handwork and reminds us on the state of software development from 50 years ago. To apply the “engineering” as a systematic methodological approach to software language definitions, we address several outstanding topics, such as lack of semantics, language libraries, compositionality, extensibility and variant management in languages and their tools. We also examine the state of the art and discuss possible solutions.

Software is eating the world. Which means that we need good languages to write all this software. And, often, come up with “new” languages to be able to efficiently target new software domains and technologies. In this talk, we will explore the changing landscape of language engineering and discuss how Artificial Intelligence and low-code/no-code techniques can play a role in this future by helping in the definition, use, execution, and testing of new languages. Even empowering non-tech users to create their own language infrastructure. Maybe without them even realizing.

Bio: I’m the head of the Software Engineering Unit at the ITIS department of the Luxembourg Institute of Science and Technology (LIST). Previosly, I’ve been an ICREA Research Professor at Internet Interdisciplinary Institute, the research center of the Open University of Catalonia (UOC) where I led the SOM Research Lab. I was also Visiting Professor at the Western Norway University of Applied Sciences, associate professor at École des Mines de Nantes as part of an Inria International Chair and postdoc at the University of Toronto. My research falls into the broad area of systems and software engineering, especially promoting the rigorous use of software models and engineering principles in all software engineering tasks while keeping an eye on the most unpredictable element in any project: the people involved in it. Current research topics include pragmatic formal verification techniques, analysis of open source communities, open data exploitation and the role AI can play in software development (and vice versa). Let’s use all the tools at our disposal to build Better Software Faster.

Model-Driven Engineering has witnessed a technological revolution with the advent of rich web interfaces. A part of the consequences is visible in the widespread adoption of Low-/No-code development platforms. From a software language engineering perspective, nowadays many language workbenches are web-based (or provide a web-based alternative). In this talk I will discuss recent experiences with the development and adoption of a “native” web-based language workbench called JJODEL. In particular, the advanced web interfaces disclose the support for interesting MDE-relevant features, notably flexible concrete syntaxes, language co-evolution, and collaborative development.

Bio: Antonio Cicchetti is an Associate Professor in Software Engineering at Mälardalen University, Innovation, Design, and Engineering department (IDT), in Västerås, Sweden. He is co-leader of the Automated Software language and SOftware engineering (ASSO) research group, which focuses on automating the engineering of software (programming/modelling) languages and software by applying advanced computation and data manipulation. Antonio is interested in the many issues related to industrial adoption of model-based system and software engineering paradigms, notably language engineering and adoption, automation supported by model transformations, consistency management, continuous development, etc. Moreover, Antonio is interested in the design of gamification systems and digital twins.