An ontology-based shared semantic space for interacting between human and computer

The discrepancy between nature language of user and special language about certain domain is a serious clag for interacting between them, and restrain the application and intelligence of knowledge-based system. In this paper, an architecture named S3 is presented, and some semantic connectors are designed to describe the relation between users´ question ontology and system knowledge ontology. For the shared semantic space, an algorithm is designed to calculate the semantic distance of the main components of the sentence with the structure <verb>+<noun>. So, users´ questions are mapped to titles of knowledge segment stored in knowledge base, and are knowledge segment as answers ranked according to the semantic similarity. Compared with existing method, this method is more feasible, and can be used in human-machine system to search answer of users´ question in knowledge-based system directly.