Flemish expertise indispensable in development of new class of AI systems
Twenty leading organisations across nine countries have joined forces to develop a new class of AI systems. The DVPS project (Diversis Viis, Plurima Solvens) is a flagship initiative aimed at advancing multimodal foundation models (MMFM) – a pioneering domain within AI research. Flemish research organisation VITO has been given a key role in this unique project, focusing on health, geo-AI, and the ethical application of AI.
In recent years, the European Union has significantly increased its investments in artificial intelligence (AI). This is the only way for Europe to safeguard its technological sovereignty, ethical leadership, and global competitiveness. With €29 million in support from Horizon Europe, the DVPS project is one of the largest European investments in AI research to date. VITO is contributing to DVPS in two crucial areas: cardiology (health) and geo-intelligence, as well as playing a role in creating explainable and transparent AI systems.
Cardiology
In the health domain, VITO plays a central role in the development of privacy-preserving, federated multimodal models tailored to cardiological applications. These models integrate various data modalities – from medical imaging and electrocardiograms to biochemical profiles – to support clinical decision-making and enable cost-effective, ethical diagnostic innovation. The data-driven, multimodal decision support improves diagnostic accuracy without compromising patient privacy.
Geo-intelligence
In the field of geo-intelligence, VITO contributes through multi-resolution environmental modelling and AI-based disaster management. By integrating multispectral satellite, aerial and drone imagery, 3D LiDAR, GIS layers, and textual sources such as social media into multimodal models, VITO enables real-time decision-making in crisis management scenarios. Through enhanced Earth observation capabilities, it supports climate resilience, urban sustainability, and environmental protection.
Transparency and Trust
Within the DVPS project, VITO is also responsible for promoting transparency and trust in AI systems through explainability, bias evaluation, and ethical alignment. This includes developing methodologies and tools to ensure DVPS models comply with EU standards for trustworthy AI, in line with the AI Act and GDPR. By incorporating ethical compliance, scientific rigour and modular design, DVPS not only strengthens Europe’s AI capabilities but also sets a global benchmark for responsible innovation.
Strategic Turning Point
For Flanders and VITO, DVPS marks a strategic turning point. It not only strengthens the region’s ambitions in digital transformation and AI-driven societal innovation but also places VITO at the heart of a pan-European effort to make foundation models more scalable, interoperable, and human-centred. Participation in DVPS reinforces VITO’s leadership in responsible AI while supporting Flanders’ regional goal of translating scientific excellence into societal and economic value.
About DVPS
The DVPS founding team consists of 70 leading European AI scientists from 20 prominent organisations:
Research: VITO (Flemish Institute for Technological Research), University of Oxford, Alan Turing Institute, École Polytechnique Fédérale de Lausanne, ETH Zurich, Imperial College London, Fondazione Bruno Kessler, Karlsruhe Institute of Technology, and Universitat de Barcelona.
Specialised partners: Heidelberg University Hospital, Vall d'Hebron Institute of Research, Amsterdam University Medical Centre, Deepset, Sistema, MEEO, Lynkeus, Data Valley, and Pi School of AI.
High-performance computing (model training): Cyfronet, the national HPC centre of Poland.
