The ontology learning from text cycle consists of the consecutive phases of term, synonym, concept, taxonomy and relation extraction. In this paper, a proposal towards the unsupervised relation extraction method is presented. Our approach is based on text documents and a set of domain axioms which represent the requirements on concepts and relations the user is interested in. We propose to use a feedback cycle that relates linguistic, statistical information and domain axioms. The evaluation is conducted on a corpus describing academic events. However, we believe that the methodology can be applicable in other domains as well. The lessons indicate that the approach is complementary to supervised relation extraction methods and can be used in conjunction with them as a mean to bootstrap an initial ontology.