ľ¹Ï¸£ÀûÓ°ÊÓ

Knowledge graphs are used to represent structured content using entities and relations on the web. To date, there exist a number of knowledge graphs (KGs) such as Wikidata, DBpedia and the Google Knowledge graph. Despite their popularity, most of existing KGs focus mainly on static data thus impeding to analyse and derive new timewise knowledge. KGs such as DBpedia allow to query the latest version of data only, although making available historical data in the form of datadumps.  On the other hand, while Wikidata maintains a small amount of temporal data, it does not provide a temporal query language to access such data.  In this case, the design of a data model to store data versions as well as infrastructures for query processing is left open.
In this talk I will describe my ongoing research activities on temporal KGs. First, I will present data models for temporal KGs which are pivotal for the efficient maintenance and retrieval of (temporal) knowledge. Second, I will introduce new query languages, which differently from existing proposal, is compatible with standard KG query processing engines. Third, I will point out the challenges in data management and querying of timewise knowledge.

About the speaker

Mel Chekol is an assistant professor at Utrecht ľ¹Ï¸£ÀûÓ°ÊÓ where he is a member of the Data Intensive Systems (DIS) group at the department of information and computing sciences. Before joining DIS, he worked as a senior researcher at the Data and Web Science group of the ľ¹Ï¸£ÀûÓ°ÊÓ of Mannheim, Germany. Mel's research interests include: static and dynamic knowledge graphs (e.g. reasoning, completion, maintenance, construction), knowledge discovery, query analysis and optimization, scalable probabilistic inference, and ontology based neural symbolic reasoning and learning.