About the Serendipity Engine

Serendipity describes experiencing unexpected, pleasant and favourable events. It involves chance encounters, coincidental events and all kinds of spontaneous discoveries. Serendipity can lead to attending a concert by an artist you didn't know, taking a shortcut through an unfamiliar neighbourhood in the city, or taking a picturesque detour on your way home. Serendipity in the city brings joy, connection and inspiration to citizens, commuters and tourists alike.

The work from Imec (IDLab) focuses on enabling automated and serendipitous discovery of data on the Web. It looks into how Linked Data and Semantic Web technologies can support applications to automatically understand explicit (and implicit) semantic descriptions of data and metadata (e.g., based on the W3C DCAT standard) to find and select relevant data sources that lead to serendipitous query answers. Concretely, the work aims to research how data publishing techniques such as Linked Data Event Streams can be applied and extended with richer semantic descriptions (e.g., using SHACL for schema descritptions) to support decentralised query processing techniques such as Linked Traversal-based Query processing. Currently this reasearch ideas are being explored on rich datasets such Publiq's Uitinvlaanderen and OpenStreetMap data sources, among others.

Adrem’s contribution to the Serendipity Engine centers on the technological core: recommendation algorithms. Within the project, we investigate how different algorithmic approaches can either foster serendipitous discoveries on personalized platforms or, conversely, contribute to harmful filter bubbles that inhibit such experiences. We also identify and examine inductive biases that may limit users’ encounters with serendipity.

Imagine exploring your city by bicycle, but without following a predetermined route or heading to a specific destination. Instead, you follow a path that leads you past picturesque streets, small parks and unknown sights. This is exactly what we want to investigate with the imec research team by developing a proof-of-concept for a cycle route planner. This route planner will not show the fastest route from point A to point B, but rather the most pleasant one, with a number of elements that can be personalised depending on the user. People who love culture can be diverted past art objects, while those who prefer nature can be guided along smaller, quieter paths. Imec will take the lead in this.

Imec is also responsible for coordinating the various pilots, the development of designs, and the coördination of the user tests.