Let’s Make the most out of INSPIRE


This project and the associated applications were created by Minerva Intelligence as part of the 2019 INSPIRE Helsinki data challenge. This event was broken into 4 different challenges, and we chose to participate in the “Let’s make the most out of INSPIRE” data challenge. This data challenge was hosted in partnership with the JRC. Our submission is a working application along with codelist and schema extensions, which enabled us to overcome limitations in INSPIRE and make it work.

The Team

Minerva Intelligence Inc. (“Minerva”) is a public company listed on the Toronto Venture Exchange. We are based in Vancouver, British Columbia, Canada, and consider ourselves leaders in the integration of semantic standards into artificial intelligence applications.  We have recently established a subsidiary company, Minerva Intelligence GmbH, in Darmstadt, Germany.

Learn more about Minerva

Left to Right: Sharon Lam, Jake McGregor, Blake Boyko, Gio Roberti, Bryan Barnhart, Clinton Smyth

INSPIRE Data Use Case

INSPIRE provides datasets with standard data structures and consistent vocabularies, enabling reusable and scalable AI applications across technical fields and regional jurisdictions. INSPIRE-aligned data is amenable to use in AI systems because analytical techniques and workflows can be re-used with minimal customization. By incorporating the use of code lists, INSPIRE data is interoperable on the semantic level, as well as in its data structure. Semantic interoperability is critical because Minerva’s AI system reasons with knowledge and data structured into semantic networks. Minerva’s system can then generate similarity rankings between expert-defined conceptual models (human knowledge) and INSPIRE aligned data. The application developed for this data challenge compares spatial data to conceptual models of different landslide types: debris flows, slides in soil and slides in rock. By aligning the semantics of our conceptual models and cognitive reasoning system to the terminology standardized in INSPIRE code lists, we demonstrate that INSPIRE data is not only interoperable for data exchange, but that it is particularly appropriate for use by powerful artificial intelligence applications.  

The use of standardized vocabularies in INSPIRE code lists enables data analysis in complex cognitive AI systems. To demonstrate this use case, Minerva created landslide susceptibility and hazard maps in the Veneto region of Italy.

Project Workflow

01

Data Collection

Select data sets were then aligned to INSPIRE standards paying specific attention to themes with hierarchical code lists.

02

Extend INSPIRE

Where necessary, INSPIRE standards were formally extended.

03

Spatial Analysis

The study area is subdivided into slope units and stream buffer polygons. These polygons are assigned values from the input data using spatial overlay analysis to create unique “instances”.

04

Reason

Instances are compared against landslide models using Minerva’s reasoning engine to calculate the relative spatial likelihood of occurrence.

05

Map

The results are displayed in an interactive web application and aligned to INSPIRE using the Natural Risk Zone Susceptibility Extension.