|
|
Evaluation of chat bots
Hazdra, Jan ; Sigmund, Tomáš (advisor) ; Čermák, Radim (referee)
This thesis deals with the evaluation of theability of machines to imitate human behavior. Presents a brief history of the development on the field of artificial intelligence and intelligent machines, specifically chat botsand describes their evolution throughout the 20th century. It describes the most important features and building blocks of chat bots, including the most widely used programming languages for creating them.It also describes the theory of A. Turing, the test and the imitation game, on which is based methodology for evaluating chat bots. Then it also examines the criticism of the Turing's test by J. R. Searle's chinese room argument that is used for its better understanding and finding its boundaries. The work also describes H. Dreyfus's assumptions, which led to optimism in the late 90s in development in artificial intelligence and limits of artificial intelligence, on which he pointed out. The main objective is the metodology proposed in the second part, which is used for comparing selected chat bots and then used for their evaluation. Overall evaluation consists of evaluations in the sub-categories in which will be individual chat bots scored. The methodology is based on the Turing test, which slightly modifies, but also extends thanks to the ideas of other authors.
|
guest ::
