Classic 20Q

This is not the first computerized version of the "what am I thinking of?" party game 20 Questions (see Animal Vegetable Mineral and the mainframe predecent Animals) but it is likely the most comprehensive, boasting a neural net database that's been cultivated since 1988.

How it works: the player thinks of some object or noun, and defies the computer to guess what it is by asking questions -- classically up to 20, but optionally even more. Rather than asking simple "yes/no?" questions, the computer asks the player to rate its questions on a spectrum of relevance including Yes, No, Unknown, Irrelevant, Sometimes, Probably or Doubtful. This not only leads the algorithm to the player's hidden focus, but also teaches it. Ultimately, it guesses what the player is thinking of, unless the player has been feeding it bad input, in which case it takes a bit longer.

A bevy of theme-specific spin-offs of the technology have been branched off (the original all-purpose database rebranded from 20Q to "Classic 20Q") and it has been exported to portable dedicated electronic devices.