How research and practice come together at Fontys

In the latest episode of the AI Garage Podcast, Erdinç Saçan talks to researchers Job Vogels and Ayrton Wouters about the role of artificial intelligence in practice-oriented research. Drawing on their work at Fontys, they give listeners an insight into the ART-IE project, in which AI and innovation come together to help SMEs adopt smart technology. Drawing on their backgrounds in artificial intelligence and logistics, they share insights into how data, collaboration and new AI applications contribute to more efficient and future-proof processes.

AI in practice: from hype to practical value

Job and Ayrton explain how companies are often interested in AI but don’t really know where to start. The hype surrounding generative AI has sparked a great deal of curiosity, but in practice, the greatest value often lies in analysing existing data. Through machine learning, companies can make processes more efficient, reduce costs and operate more reliably. Particularly in sectors such as logistics, where large amounts of data are available, AI offers opportunities to optimise complex processes such as planning and decision-making, whilst more standardised processes such as order processing benefit primarily from further automation and error reduction.

Collaborating without sharing data: the power of federated learning

A key component of ART-IE is the use of federated learning: a form of AI in which models learn from data without the need to share that data. This enables collaboration between organisations without privacy or competitive risks. Companies train their own models locally and share only the insights, creating a collectively smarter system. This is particularly valuable in sectors where data is sensitive. At the same time, it is emphasised that a successful AI application starts with good data. It is important for an organisation to consider a ‘roadmap’ of changes needed to successfully implement smart data solutions within the organisation. Many projects begin with cleaning and structuring data; only then does real AI development become possible.

Listen to the full episode of the AI Garage Podcast via Spotify or Apple Podcasts (in Dutch).