s-X-AIPI open-source toolset is now live
Over the course of the project, the s-X-AIPI team developed a set of open-source tools designed to help process industries adopt more trustworthy and self-X AI solutions.
These contributions — now publicly available on platforms like Zenodo, GitHub, and the AI-on-Demand (AIoD) platform — focus on different industrial applications, particularly in steel, aluminium, asphalt, and pharmaceuticals.
What’s Included
The open-source collection includes AI toolkits, software, models, datasets, and documentation, all structured to help industry partners, researchers, and developers effectively work with self-X AI systems.
AI Toolkits & Libraries
At the core of the toolset is the Autonomic Manager and a self-X package built in R. These provide foundational tools and libraries for building self-X AI solutions with autonomic capabilities like data exploration, anomaly detection, feature selection, and self-optimization — all essential for systems that need to adapt on the fly.
AI Models
We also developed and released specific models tailored to industrial use cases that address tasks like:
Prediction of burner power
Forecast of aluminium composition
Metadata generation
Fine-tuning and retraining existing AI systems
Datasets
Several datasets were shared alongside the models, especially those used for model training or metadata generation for the interaction with the Autonomic Manager.
Supporting Materials
Documentation includes data management resources, software components, and templates — such as a sustainability data model designed for the aluminium sector. A research paper outlining the self-X pipeline for steel production is also publicly available.
Why It Matters
These open-source tools represent a major technical output of the project and one of its most impactful. By making them publicly accessible, we hope to support further experimentation, replication, and application in industrial environments. This also aligns with the broader EU goal of fostering open, responsible AI development across strategic value chains.
You can explore the resources now on:
