Framatube

Researchers, educators, and developers face a common challenge when building robust, ethical, and sustainable AI software: the lack of accessible full-stack development resources and expertise. Low-code platforms promise to address this gap by automating parts of the coding process and reducing development time. But an important question remains: are these platforms truly open, or do they lock users into proprietary ecosystems?

Often, proprietary low-code tools create vendor lock-in and limit user freedom. We argue for more open-source alternatives that empower users to use, extend, and even adapt the tools to fit their specific needs.

In this workshop, we present BESSER, an open-source low-code platform for smart software development. Participants will get hands-on experience creating applications, databases, AI agents, and chatbots directly from our web modeling editor. Beyond these exercises, we will also explore the advantages of BESSER’s open-source ecosystem, which gives users the freedom to extend the platform, adapt it to their own domains, and contribute new features to the project, building and sharing their own model-driven solutions.

By Armen Sulejmani and Atefeh Nirumand Jazi

Kaustik im Vergleich, es treten an: der parabolische und der runde Hohlspiegel. Mittwoch live in der Vorlesung, aber jetzt schon für Euch online!

Links als "Gegenstand" ein Buckyball; rechts in grün (Kugel) und blau (Parabel) zwei Hohspiegel, welche unseren Gegenstand abbilden. Die entsprechenden Strahlbündel sind ebenfalls grün bzw. blau dargestellt; an ihren Treffpunkten bilden sie jeweils das Objekt ab. In weiß dargestellt ist das "ideale" Bild für eine abbildungsfehlerfreie Optik.

Die zwei Kurven geben die Kaustik an, hier im Sinne: an welcher Stelle das Bild unsere Gegenstandes entsteht, abhängig von dessen vertikalen Position. Ideal wäre die längentreue Abbildung in eine Ebene.

Most AI today runs on giant, opaque, black-box models — this talk explores how open source language models offer a sovereign, transparent, and responsible alternative.

By Benjamin Bellamy and Andrzej Neugebauer

This presentation offers a concise overview of the current state of the art in AI and robotics, highlighting key developments, capabilities, and limitations of modern systems such as ChatGPT and its counterparts. We will explore the growing societal and technological fascination with AI-driven agents and machines, and what this means for the future of human-machine interaction. From breakthroughs in large language models to advances in autonomous systems, this talk invites reflection on how far we've come—and where we might be heading next.

By Serge Linckels

The rapid adoption of AI-assisted coding tools like GitHub Copilot and ChatGPT has accelerated software development processes, but it has also introduced significant risks. Developers may unknowingly use AI-generated code that violates licensing restrictions or includes vulnerable third-party dependencies.

In this talk, Philippe will share a new approach, using open source tools and open data, to identify and locate AI-generated code in software projects and products for safer, efficient, and more responsible and ethical use of AI-generated code.

By Philippe Ombredanne

The (AI-Horde)[https://aihorde.net/] demonstrates that crowd sourcing computing power for generative AI is viable. Our volunteers provide their resources for overwhelming altruistic reasons, with many believing that generative AI has profound educational and artistic benefit. By providing a platform for those volunteers to offer their compute in a fair and equitable way using our ‘kudos’ system, we have demonstrated that significant ideological will exists to provide these services for free to the public.

By Konstantinos Thoukydidis

This talk presents Gospel (Generic OCaml SPEcification Language), an emerging open source ecosystem for accessible formal verification currently under development through French National Research Agency funding. Gospel enables developers to write strongly-typed behavioral specifications that can be consumed by multiple verification tools.

While the work on static verification tools like Cameleer is in progress, the ecosystem demonstrates a "gradual verification" approach where organizations can start with dynamic testing of all code and progressively apply formal proofs to critical components. The Ortac tool, developed at Tarides, translates Gospel specifications into executable OCaml code for defensive programming, runtime assertion checking, and specification-driven testing.

The ecosystem emerging around Gospel provides a practical entry point to formal methods that industry can adopt today!

By Sabine Schmaltz

The CRA is already here, and will impact software development worldwide. Organizations need efficient compliance processes, supported by excellent free and open source tools to correctly identify and manage software components and secure software supply chains.

Join us for the latest updates on how to use free and open source tools, data, and standards for better, faster, more efficient, and automated software supply chain management, for public and private organizations of all sizes.

By Philippe Ombredanne

Incident response and threat intelligence teams often face a paradox: vast information but no structured way to manage cases, tasks, and collaboration, leading to duplication, silos, and delays.

By David Cruciani