Explaining Semantic Reasoning Using Argumentation

Multi-Agent Systems (MAS) are popular because they provide a paradigm that naturally meets the current demand to design and implement distributed intelligent systems. When developing a multi-agent application, it is common to use ontologies to provide the domain-specific knowledge and vocabulary necessary for agents to achieve the system goals. In this paper, we propose an approach in which agents can query semantic reasoners and use the received inferences to build explanations for such reasoning. Also, thanks to an internal representation of inference rules used to build explanations, in the form of argumentation schemes, agents are able to reason and make decisions based on the answers from the semantic reasoner. Furthermore, agents can communicate the built explanation to other agents and humans, using computational or natural language representations of arguments. Our approach paves the way towards multi-agent systems able to provide explanations from the reasoning carried out by semantic reasoners.