Statistics and Risk Management Seminar - 6/5/2015

Wednesday, 6 May 2015, 2:00 p.m.

Lecturer: Professor José Alferes, NOVA LINCS, Departamento de Informática, FCT-NOVA

Title: "Web of Data"

Local: Room 1.3, Edifício VII

Faculdade de Ciências e Tecnologia, Quinta da Torre, Caparica

Abstract: What if the Web could be used as a huge database? What if one could make complex queries using and relating data that exists in several diff erent web portals?

One could, for example query the web for the capitals of the birth countries of actors participating in series broadcasted on a given TV channel last night. Google wouldn't help answering this. But of course you can get the answer by going to the web pages of the channel, for knowing which series were broadcasted, then to imdb, to know who are the actors in those series and possibly their nationalities, and then to wikipedia to know the capitals of those country. This would be a tedious task, but feasible anyway. But imagine that you wanted to know, for each country, how many actors were born there that ever participated in a TV series broadcasted on weekends. Would you still be able to do it in this way? And what if you could do it as in a database, with a single query?

You may rightly say that the example is dummy, and no one would ever want to do such a query. But imagine that you could query the web, just like that, using data in portals of airline companies, hotels, booking agencies, weather information, public transports, strikes, wikipedia, theatres, museums, etc. Wouldn't that be great for preparing your holidays? Or imagine that you could query the web, combining data from governamental institutions, with Pordata, with public data on the accounts of companies. Or if you could mix, in a single query, data from PubMed, with di fferent databases with all kinds of knowledge about life sciences. 

Today, much of this is becoming possible. In fact, an infrastructure and tools are being built that allow one to use the Web, not as a huge amount of interconnected documents that you can read (and index to fi nd text), but as a huge amount of interconnected data, that you can query and process automatically. The construction of such a Web of Data, not only opens big opportunities, but also raises many technological and scienti c questions. In this seminar I'll let you know about this Web of Data, the research challenges it faced (including in areas close to mathematics), and the research challenges that still lie ahead.