Skip to main content

November 25, 2015

Speaker: Mehdi Kargar, Mitacs Elevate Postdoctoral Research Fellow

Title: Keyword Search in Structured Big Data

Abstract: Huge amount of data is generated every day. In fact, the digital universe is doubling in size every two years. The effect of big data to improve our capabilities and lives is limited to our ability to use the data. Having a strong system that enables users to access big data is of paramount importance. Much of the world’s high-quality big data are stored as structured data. This includes enterprise’s RDBMSs, XML repositories and social networks’ big graph (i.e. the connections among people in social networks form a big graph). In order to use the structured data, one should be familiar with its structure (i.e. schema) and a query language (e.g. SQL). However, the complexity of query languages and database structures has largely restricted its use to experts and skilled developers. In contrast, a non-technical end-user is effectively locked out. Such an end-user is limited to use pre-defined forms to use the data.

Keyword search over structured data (and its variations in social networks) offers an alternative way to access and use structured data that neither requires mastery of a query language, nor deep knowledge of the database’s potentially quite complex schema. I will talk about problems, challenges and opportunities for improving structured data exploration. More specifically, my work revolved around building systems to resolve these three problems: keyword search in big graphs, keyword search in relational databases with complex schema and team formation in social networks. I will also briefly talk about my experience with the industry and how theory can be implemented in real-world systems.