Induction in Nonmonotonic Causal Theories for a Domestic Service Robot
Jianmin Ji, Xiaoping Chen · IMPERIAL COLLEGE PRESS eBooks · 2014
Abstract. It is always possible to encounter an expected scenario which has not been covered by a certain theory for an action domain. This paper proposes an approach to treating this problem. We reduce this learning task into the problem of modifying a causal theory such that the interpretations corresponding to new scenarios become a model of the updated theory, while all the original models keep unchanged. We illustrate our approach through a case study based on a domestic service robot, KeJia. 1