Geospatial Insights - Unveiling Urban Diversity

Looking at the application of AI to discover the location of Roman roads in the Scottish Borders, the talk will start with a brief explanation of LiDAR data and then show how it can be applied to the discovery of previously unknown Roman roads.

User-generated, online data is rich with insights into our world. Interpreting the spatial information tying this data to real-world locations is a key challenge. Modern advancements in AI and machine learning might provide a promising solution – but we first need to establish just how well these models really understand spatial information. To answer this, we will explore the ambiguities and complexities of geospatial language, and the limitations of modern AI models.

It is often assumed that the “majority rules”, but is this actually the case? In many instances, the way a system works is determined by a small, inflexible group (sometimes called a “stubborn minority”) who—for various reasons—may not (or, perhaps, cannot) adapt. This talk outlines what is meant by this term, considers examples of “stubborn minorities” in cities, and discusses why this matters for everyone.