Knowledge Representation and Reasoning: What’s Hot

Knowledge representation and reasoning (KR) stems from a deep tradition in logic. In particular, it aims at build-ing systems that know about their world and are able to act in an informed way in it, as humans do. A crucial part of these systems is that knowledge is represented symbolically, and that reasoning procedures are able to extract consequences of such knowledge as new symbolic representations. Such an ability is used to deliberate, in an informed fashion, about the course of actions to take. This very idea is radically new in human history (Levesque 2014). It comes about after a long gestation, stemming from Aristotle, who developed the initial notion of logic though unrelated to the notion of computation; con-tinued by Leibniz, who brought forward a notion of “think-ing as computation, ” though not yet symbolic; and later by Frege, who developed the notion of symbolic logic, though unrelated to computation; and finally by the breakthrough in human thinking of the early part of last century with Church, Godel, and Turing, who set the bases for symbolic logic bound together with computation and ultimately for Com-puter Science, though even they did not think about logic as a way of representing knowledge. It was McCarthy (McCarthy 1959) who first had the very radical idea of building intelligent systems by focusing not on programming or on the details of the system architecture, but on the knowledge necessary to behave in an intelligent way. If a system were able to represent what it needed to know and draw conclusions from it in an automatic way, it would be able to deduce for itself how to behave intelli-gently. Knowledge representation and reasoning is a well-established field of AI deriving directly from McCarthy’s idea. Research in KR is currently present in the major AI generalist conferences like the International Joint Confer-ence on Artificial Intelligence (IJCAI), the AAAI Confer-ence on Artificial Intelligence (AAAI), and European Con-ference on Artificial Intelligence (ECAI)