Cognitive true reflection of characterization, simulation calculations are symbolic imitation simulation. The essence of the algorithm is based on rational thinking based on computational logic, the lack of non-facts or counterfactual imagination process, namely the fact that object-oriented, attributes, relationships, constantly changing cognitive dynamic adjustment process.
Data will not tell you, but cognitive - will.
Q: Are you suggesting that there is a trend away from the machine learning it?
Judea Pearl: not a trend, but a serious self-reflection, including these questions: Where are we going? What's next?
Q: That is the last thing I want to ask you a question.
Judea Pearl: I'm glad you did not ask me questions about free will.
Q: So, what do you think of free will? (Intelligent Bacteria mark: so embarrassed, as the questioner wit me ~)
Judea Pearl: We will definitely develop a robot with a free will. We must understand how to program them, and from what we can get. For some reason, evolutionary speaking, this free will can be calculated.
Q: In what way?
Judea Pearl: You have free will, evolution makes us feel that way. Clearly, it provides some computing functions.
Q: When a robot appear free will be very obvious?
Judea Pearl: I think the first evidence that, if the robot starts counterfactual communicate with each other, such as they say, "You should do a better job." If a robot team began to communicate in this language, then we will know that they have a free will. "You should give the ball to me - I'm waiting for you, and you do not!" "You should do something" means that you can control to do or not do something. So the first sign of the exchange, the second is the ability to play better football.