| Project name: | European Large Open Multi-Modal Foundation Models For Robust Generalization On Arbitrary Data Streams |
| Acronym: | ELLIOT |
| LinkedIn profile: | www.linkedin.com/company/elliot-eu/ |
| ID code: | 101214398 |
| Call: | HORIZON-CL4-2024-HUMAN-03-01 |
| Supported by: | European Union, Horizon Europe |
| Project duration: | 07/2025-6/2029 (48 months) |
| Principal investigator at CIIRC: | Dr. Ing. Josef Šivic |
| Coordinator: | Ethniko Kentro Erevnas kai Technologikis Anaptyxis, Greece |
| Project partners: | Eberhard Karls Universität Tübingen, Germany Forschungszentrum Jülich GmbH, Germany Universiteit van Amsterdam, Netherlands Technische Universiteit Eindhoven, Netherlands Università degli Studi di Trento, Italy Centre de Visió per Computador, Spain Institut Jožef Stefan, Slovenia Globaz, S.A., Portugal Barcelona Supercomputing Center – Centro Nacional de Supercomputación, Spain CSC – IT Center for Science Ltd., Finland CINECA Consorzio Interuniversitario, Italy Ludwig-Maximilians-Universität München, Germany Universitat de València, Spain Università degli Studi di Modena e Reggio Emilia, Italy Aalto University Foundation SR, Finland Fundación de la Comunitat Valenciana Unidad ELLIS Alicante, Spain ELLIS Institute Tübingen gGmbH, Germany Czech Technical University in Prague, Czech Republic CISPA – Helmholtz Center for Information Security gGmbH, Germany Katholieke Universiteit Leuven, Belgium Voxist, France Valeo Comfort and Driving Assistance, France Robotwin s.r.o., Czech Republic Openchip Software Technologies SL, Spain DEIMOS Engineering and Systems SLU, Spain DEIMOS Space UK Limited, United Kingdom DEIMOS Space Sociedad Limitada Unipersonal, Spain Departament de Territori, Habitatge i Transició Ecològica, Spain De Vlaamse Radio en Televisieomroeporganisatie NV, Belgium École Polytechnique Fédérale de Lausanne, Switzerland Eidgenössische Technische Hochschule Zürich, Switzerland |
| Budget: | Total budget: € 24,998,023.75 EU support: € 24,998,023.75 Support for CIIRC CTU: € 937,500.00 |
| Media and Press Releases: | 17.6.2025: ELLIOT: A Flagship Initiative to Develop Open Multi-modal Foundation Models for Robust Artificial Intelligence in the Real World |
Abstract:
The ELLIOT project aims to develop the next generation of Multimodal Space-Time Foundation Models (MSTFMs), significantly enhancing the capabilities of general-purpose AI models in domains where the temporal dimension plays a critical role. These models will integrate spatio-temporal understanding with support for modalities not yet commonly addressed in large foundation models – such as industrial time series, remote sensing data, and health-related measurements.
While some modalities like visual input are already present in today’s generative AI systems, MSTFMs will go further – offering a unified framework for processing time-dependent data across various domains. A combination of real and synthetic data will be used for training. Real-world data will be sourced from consortium partners and European data spaces, while synthetic data will be generated using both existing generative AI techniques and new methods developed within the project.
A key priority of the project is European technological sovereignty. The necessary computational infrastructure for training these large models will be ensured through European High-Performance Computing (HPC) resources included directly in the consortium. This will allow the MSTFMs not only to be developed, but also deployed in practice – in areas such as healthcare, industry, and environmental monitoring.
The project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No. 101214398, within the topic HORIZON-CL4-2024-HUMAN-03-01 “Advancing Large AI Models.”
