The Binary Tree Structure
Assuming that each question that the Akinator asks represents a node in the neural network, the most appropriate method to get the correct answer within 20 questions would be eliminating roughly half of the possibilites within each question. A typical starting question for the Akinator is "Is your character real or fictional?". This question will eliminate roughly half of the possibility pool by knowing if the character is in the pool of fictional or non-fictional characters, thus reflecting a binary tree. An example of a binary tree is shown below. Although improbable, if every question can eliminate half of the possibilities, the most popular 1.05 million (2 to the 20th power) will be successfully found within 20 questions. In conclusion, the neural network is programmed to emphasize question that will typically split its possibility pool in half for the most efficient for of question-asking.
Assuming that each question that the Akinator asks represents a node in the neural network, the most appropriate method to get the correct answer within 20 questions would be eliminating roughly half of the possibilites within each question. A typical starting question for the Akinator is "Is your character real or fictional?". This question will eliminate roughly half of the possibility pool by knowing if the character is in the pool of fictional or non-fictional characters, thus reflecting a binary tree. An example of a binary tree is shown below. Although improbable, if every question can eliminate half of the possibilities, the most popular 1.05 million (2 to the 20th power) will be successfully found within 20 questions. In conclusion, the neural network is programmed to emphasize question that will typically split its possibility pool in half for the most efficient for of question-asking.