At the heart of Turing's 1950 imitation game is the question-answer test to assess whether a machine can respond with satisfactory and sustained answers to unrestricted questions. In 1966, Weizenbaum's Eliza system made it possible for a human and a machine to communicate via text in question and answer sessions. Forty five year later in 2011, IBM Watson achieved remarkable success winning an unrestricted question-answer exhibition match competing against humans in Jeopardy! a US TV quiz show. Is it now time to scale up to Harnad's Total Turing Test combining natural language with robot audio and vision engineering?
展开▼