The Study Of English Verb Clustering Based On Human Robot Interaction

How to enable service robots to understand their user intention is a hot topic in research today, based on the understanding, the robots can coordinate and adjust their behaviours to provide desired assistance and services to the users as capable partners. Active robot learning (ARL) is an approach to the development of the understanding of humans' intention. Task action bank is one part of the ARL which can store task categories. In this approach, a robot actively performs test actions in order to obtain its user's intention from his responses to the actions.This paper is introducing an approach to classify the verbs based on the basic action of robot using statistical method. A parser will be established use for treat with corpus and analyze the probability of verb's feature vector.When it comes to extract the verb vector, a series algorithm has been employed to accomplish. Firstly, we define four basic task categories for service robot. Secondly, we pick up the usual noun through the WordNet which is a representative of relationship between words. Then, we calculate the semantic relatedness of verb and noun by PMI and establish the bipartite graph and apply WJ and k-mediod to complete the cluster.Experimental results will present the collocation between the similarity verbs and it could be combining with artificial intelligence to establish test action bank for ARL (Active Robot Learning).