Winograd Schhema Challenge Essay

There are other tests which are an alternative to the Turning test, to test whether a computer can have a conversation like a human. Hector Levesque, a Computer Science professor at the University of Toronto states that “The Turing Test relies too much on deception. A computer program passes the test if it can fool an interrogator into thinking she is dealing with a person not a computer”2. So he proposed the Winograd Schema Challenge, this test consists of a specifically styled question that has multiple choice, as an answering format. This test uses simple questions that can be answered by any human, but unlike the Turing test, these questions require logic, common sense, and reasoning that can only be fulfilled with true human intelligence.

When actually they have just been programmed with certain abilities, and they cannot extend beyond this, unlike us human, who can develop and extend their knowledge, and logic. Levesque’s basic theory behind this Winograd Schema Challenge, is testing the chatterbot by asking it a question that it is likely not to be familiar or programmed to have an answer to, and so it tests the chatterbot’s immediate response based on its pure intelligence. Lovesick published another sample question which displayed this “Could a crocodile run a steeplechase?”2 this question is so unique, many of us human may not have never heard or thought of it, but when we hear this question an answer forms in our mind due to the logical, and common intelligence that we have, where as a chatter bot may not be able to answer this

To experiment with the techniques that chatterbot’s use to act like a human, I made conversations with three chatterbots; Alice, Cleverbot, and Eliza. Cleverbot, unlike many other chatterbots, is not programmed with responses that can be output to the user’s questions, he instead uses the technique of learning from the users. In 2011, Cleverbot took part in a Turning Test at the Techniche festival, he was judged to be 59.3% human, where as a human who also took part in the human achieved 63.3%3. This achievement by cleverbot is due to the technique which he uses, every time he gets an input from the user he store is in his large database, and when he is asked a question by the user, he conducts a database search and uses the keywords from the input to find a perfect match from past response, to reply to the current user. Since the reply from Cleverbot is one that was one typed in by a human, his responses sound like a