YAGO2 is a huge semantic knowledge base, derived from Wikipedia, WordNet and GeoNames. Currently, YAGO2 has knowledge of more than 10 million entities (like persons, organizations, cities, etc.) and contains more than 460 million facts about these entities.

Time and location data are treated first-class citizen, the basic triple model is extended by special fields for time and location for querying. Also, we took special care to consistently attach temporal and spatial data to all facts where it is semantically meaningful and where it can be derived from Wikipedia. Unlike many other automatically assembled knowledge bases, YAGO2 has a confirmed accuracy of 95%.

I like the idea, not sure how practical the applications are without more consensus on emotion scales and shared vocabulary of terms.

Google's director of search quality discusses challenges of automation, knowledge, spam, and even politics.

PDF des sessions ajouté au site