Learning Computational Verb Rules

Computational verb rules are efficient tools totransform dynamical experiences in natural languages intomathematical expressions. Based on human experiences, we caneasily construct computational verb rules. However, due to thelimits of human perceptions, it is impossible to design accuratecomputational verbs based on human perceptions. To calibratethe parameters of computational verbs in computational verbrules, we need to train them by using measurements that aremuch more accurate than human perceptions. Different kindsof learning algorithms of computational verb rules, of whichthe consequents and antecedents are static or dynamical, arepresented.