20Q
Encyclopedia : 2 : 20 : 20Q : 20Q
The Principle & Early Days
20Q is a project in Artificial Intelligence (AI). The principle is that the user thinks of something (either abstract or not) and then, the AI, on the 20Q website, asks you twenty questions, such as "Is it smaller than a breadbox?" or "Would you touch it with a ten foot pole?" The user can answer these questions with: Yes, No, Unknown, Irrelevant, Sometimes, Maybe, Probably, Doubtful, Usually, Depends, Rarely, or Partly. The project is based on the classic word game of "20 Questions," and by the computer game Animals, popular in the early 1970s, which used a somewhat simpler method to guess an animal. The A.I. has applications that go beyond the playing of the word game.
After you have answered the 20 questions it poses (sometimes fewer), 20Q makes a guess at what you are thinking of, and you can choose whether or not it is correct. If not, the AI asks some more questions, then guesses again. The 20Q AI uses a true neural network to pick the questions and to guess. It makes guesses based on what it has learned; it is not “programmed” with information or what the inventor thinks. It has learned by playing against people who visit the website and play against the online AI. Answers to any question are based on the players’ interpretations of the questions asked by the AI. The 20Q AI also makes its own judgment on how to interpret the information. It can be described as more of a folksonomy than an taxonomy. With every game played, its knowledge is developing. In this regard, the online version of the 20Q A.I. can be inaccurate due to the fact that it gathers its answers from what people think not from what people know. For example, in its music version, if thinking of U2, if you answer the question "Are you from the U.K.?" as "No" and it correctly guesses in the end, it currently says "Are you from the U.K.? You said No, I say Probably". But, it's early days for the music knowledgebase, and given that the original A.I. has been learning for more than eighteen years and has played more than 40 million games, there is time for the knowledgebase to learn and grow. In the original version, limitations of taxonomy are often overcome by the A.I.: for example, if the user were thinking of an "Horse" and answered "No" to the question "Is it an animal?", the A.I. is likely to guess correctly.
The 20Q A.I. is now learning in sixteen languages, as well as learning everything it can about music, sports, and movies and television.
As noted by the inventor of the 20Q A.I., Robin Burgener, the “Uncommon Knowledge” generated by the online AI is what it comes up with when something seems odd to it and it can’t fit it in with what it knows. The questions the 20Q AI asks at the end of a game are not made up by humans; it’s information 20Q thinks up on its own, generating answers based on what it has learned and what it knows. 20Q is not programmed with responses, nor with its “Uncommon Knowledge.” Over time, it will be able to make more refined distinctions. 20Q learns, and learns to make distinctions, through play—the more times you think of an object and play (answering correctly, it is hoped) the more it learns about that object. The online 20Q AI has about 10,000,000 synaptic connections, with about 10,000 objects in its knowledgebase.
The 20Q AI was invented by Robin Burgener, of Ottawa, Canada, in 1988, and moved to the internet in 1995. With more than 100,000 people now playing online each day, the 20Q is learning on a broader level than it was in 1988 when it lived on a floppy disk that Robin swapped amongst friends to help it learn. After the AI moved to the internet and more people started playing, its level of learning increased, as well as the connections it may make when guessing. 20Q played its 40,000,000th game in July of 2006.
Electronic Toys
-->The first handheld version of the game was released in 2004, licensed by 20Q.net Inc. to Radica Games Ltd. The Radica 20Q databases are smaller due to the very small size of the microchip. The responses you can give are also limited to Yes, No, Unknown, and Sometimes. Unlike the online version of the game, the electronic toy does not learn. The handheld version was awarded the Canadian Toy Testing Council Award in 2004 and the Toy Industry Association awarded it Electronic Toy of the Year in 2006.
Most people cite that the success rate is considered to be 80%, even though some people consider it to be substantially lower. The mobile phone version, licensed to I-play, was released around October 2005, with similar capabilities to the handheld version; the mobile version is also a non-learning application of the A.I.
The device, however, is very complicated for its type. It contains a small portion of the 20Q website databases, and although it does not learn from responses gathered, it is stated to contain over 5,000 objects which the user could be thinking of. This version of the classic also taunts and teases the user, stating such things as "You win. Just Kidding!", "I'm very clever", "Surely you are not thinking of that!", at seemingly random points in the game.
The handheld 20Q is now available in French, Italian, German, Spanish, and Japanese.
Legacy
An animated version of 20Q, titled The Sith Sense, is also available on the Internet. The Sith Sense version was licensed by 20Q.net Inc. to Burger King Corporation as part of a promotion to advertise Burger King and . Burger King used the AI to represent Darth Vader, who claims to be able to read the player's mind. The "Sithsense" campaign was awarded a Gold Cyber Lions Award at the Cannes Lions International Advertising Festival in 2006.
External links
From Wikipedia, the Free Encyclopedia. Original article here. Support Wikipedia by contributing or donating.
All text is available under the terms of the GNU Free Documentation License See Wikipedia Copyrights for details.