Research programme
Foundation for Industry (FIND) – Large AI models for a resilient high-tech industry.
Foundation for Industry (FIND) brings together 5 universities with 10 labs, 11 Dutch companies, ranging from start-ups to multinationals, and 2 knowledge institutes to pave the way for a new wave of AI-based automation that helps the Dutch industry strengthen and keep its international competitive advantage as a leading high-tech nation in the AI-era.
- Duration
- 2025 - 2031
- Contact
- Suzan Verberne
- Funding
- NWO
- Partners
University of Amsterdam / Vrije Universiteit Amsterdam / TU Eindhoven / TU Delft / TNO / NKI
FIND brings the benefits of Foundation Models (FM), i.e., large AI models, to unexplored industry-relevant applications and data types to realize a Dutch high-tech industry that is competitive in the AI-era. Foundation models are a disruptive AI technology that can simulate complex cognitive tasks and provide a basis for a new wave of AI-based automation.
FIND is set to address the risk of diminished local production capacity and increased foreign dependency, which threaten the Netherlands’ position as a leading high-tech nation. By extending foundation models to broader currently under-served (Dutch) industry-relevant applications, their full potential can be met, while at the same time enhancing economic stability and maintaining the attractiveness of the Netherlands for international talent and investments. These under-served applications require beyond state-of-the-art foundation models, with improved robustness, transparency, that can work with hitherto unexplored data modalities, and can be executed on edge devices to speed up decision-making and maximise privacy. To address this, FIND will develop a precompetitive FM Toolbox consisting of documented and open-source 1) pre-trained models, 2) required code for training and integration and 3) validated research outputs. The open-source FIND toolbox is crucial for enabling industry outside the consortium to integrate FM technology. The elements of the toolbox will be developed together with academic and industry experts from machine learning, data science and embedded computation backgrounds and validated in relevant AI use-cases.
The relevant stakeholders are involved in FIND through the User Committee, consisting of industry partners, and the Ethical & Societal Impact board. The involvement of all relevant societal, academic, and industrial stakeholders ensures the utilization of the FIND toolbox and its envisioned impact: Ensuring the Netherlands will keep its international position as a leading high-tech nation.
Involved (end)users: AMS / NXP / Canon / Shell / Technolution / ASMPT / Stryker / KAIKO / Perciv.AI / KeyGene