Learning and reasoning

Learning and reasoning are prominent targets of research in computational cognitive science. This chapter explores some of the questions in the light of recent advances in computational cognitive science such as the following two success stories. First, the Google-owned artificial intelligence company DeepMind created a system called deep Q-network (DQN) that could learn how to play forty-nine different arcade games – including Pong, Pac-Man, and Space Invaders. Second, a computational system was developed within the Bayesian framework that could learn concepts associated with handwritten characters. The chapter refers to these two cases to sketch a taxonomy of computational approaches to learning and reasoning, and discusses some implications for three issues. These three issues concern, first, the character of humans' innate cognitive architecture; second, the distinction between two kinds of thinking: one fast and intuitive, the other slow and deliberative; and third, the nature of rational behaviour.