The Ontology Web Language (OWL) for a Multi- Agent Understating System Mostafa M. Aref Zhengbo Zhou Department of Computer Science and Engineering University of Bridgeport, Bridgeport, CT 06601 email: aref@bridgeport.edu Abstract— Computer understanding is a challenge problem in Artificial Intelligence. A multi-agent system has been developed to tackle this problem. Among its modules is its knowledge base (vocabulary agents). This paper discusses the use of the Ontology Web Language (OWL) to represent the knowledge base. An example of applying OWL in sentence understanding is given. Followed by an evaluation of OWL. 1. INTRODUCTION One of various definitions for Artificial Intelligence is “The study of how to make computers do things which, at the moment, people do better”[7]. From the definition of AI mentioned above, “Understanding” can be looked as the first step for a system to realize the ability of doing things as well as humans. Natural language processing needs an understanding system to make the machine understand human languages. Understanding is a transformation from one representation to another [1]. To achieve this transformation, the input will be processed through a series of agents. From morphological analysis to pragmatic analysis, the machine can “read” the input and has its own representation. Several applications may be developed based on the understanding system. Some examples of these applications are Machine learning, machine translating, and expert systems with better performance. A multi-agents understanding system accepts a user input in a form of speech (typed or voice). Then, the user may enter several questions concerning the user input. The system should answer these questions that reflects the understanding of the input [1]. The multi-agents understanding system consists of the following agents: a morphological analyzer, a semantic analyzer, a discourse analyzer, a user interface, and a knowledge base. The knowledge base is the main module in the understanding system. It contains the English vocabulary agents and all the linguistic information about the vocabulary using object- oriented technology [1]. OWL is a Web Ontology Language. It is built on top of RDF – Resource Definition Framework and written in XML. It is a part of Semantic Web Vision, and is designed to be interpreted by computers, not for being read by people. OWL became a W3C (World Wide Web Consortium) Recommendation in February 2004 [2]. The OWL is a language for defining and instantiating Web ontologies. OWL ontology may include the descriptions of classes, properties, and their instances [3]. Given such ontology, the OWL formal semantics specifies how to derive its logical consequences, i.e. facts not literally present in the ontology, but entailed by the semantics. Section 2 describes a multi-agents understanding system. Section 3 gives a brief description of a newly standardized technique, Web Ontology Language—OWL. A working example of the OWL applied in knowledge representation is given in section 4. Section 5 evaluates the performance of OWL. Conclusion and directions of the current research are presented in section 6. 2. MULTI-AGENT UNDERSTANDING SYSTEM To understand something is to transform it from input representation into internal representation has been chosen to correspond to a set of available actions that could be performed [1]. The process of natural language understanding is as follows [7], as shown in Figure 1. Morphological Agent: given the input text, morphological analyzer converts the text into group of words in the basic form and their linguistic information. It also separates the affixes from the input tokens [1]. Semantic Agent: structures are created to represent meanings of a group of words (sentence). In other words, a mapping is made between the input sentence and objects in the task domain. Discourse Agent: Given the agent sub-societies of set of sentences, discourse analyzer agent resolves references between these sentences. The user interface is needed to facilitate the communication between the understanding system and the user [1]. For example, a web page containing several text input boxes can get input from a human and then gives another page or dialog box with the answer or some other actions.