“Concept Game is a simple yet immersive game where a slot machine generates random assertions and you decide if they make sense or not. Each meaningful assertion is a winning configuration and you score points for detecting such assertions. At the end of each month top three players receive small gift checks from amazon.com.
The interesting aspect of Concept Game is that it’s another example of the games-with-a-purpose approach (http://en.wikipedia.org/wiki/Game_with_a_purpose). While people play the game, they also help a group of researchers to annotate a data set of common sense assertions. These assertions are automatically extracted from web-based text by text mining algorithms and the game helps them to filter out the meaningless ones.
But what is common sense? It is not found in encyclopedia, it is not explicitly stated in verbal communications. It is, after all, the very set of concepts and knowledge that we assume everybody else has; stating the obvious would be redundant. Without it, we wouldn’t be able to function in everyday life and neither artificial intelligence systems can. Some studies estimate that the amount of the commonsense knowledge that a layperson posses can be represented by millions of assertions (http://web.media.mit.edu/~push/ConceptNet-BTTJ.pdf). Representing such knowledge is essential if we want to have “intelligent” softwares that can “understand” our intentions and act accordingly.”