Drawing smart conclusions from dispersed information

When a streaming service suggests a new film for you to watch, it has often explored what you usually like to watch using knowledge graphs. Wallenberg Academy Fellow Olaf Hartig will develop methods for integrating knowledge graphs, allowing smart conclusions to be drawn from information stored in different places, like different companies. 

The knowledge graphs used by streaming services to provide you with intelligent film recommendations are based on information such as the genres, actors, or directors you tend to like. The knowledge graph finds correlations in the data set, analyzes the correlations, and then searches for a recommendation for you.  
Similarly, knowledge graphs are used by truck manufacturers who need suggestions for maintenance plans for different vehicle models, for example, or by people who want to develop artificial intelligence that can answer questions in a smart way.  

To make knowledge graphs even more useful, Dr Olaf Hartig from Linköping University is developing methods for integrating separate knowledge graphs from organizations such as companies, healthcare units or authorities. The integration component will be able to ask each knowledge graph questions and then summarize the results, as if all the separate knowledge graphs were combined into one. 

Photo Patrik Lundin