Capability Modeling of Knowledge-Based Agents for Commonsense Knowledge Integration

摘要: Robust intelligent systems require commonsense knowledge. While significant progress has been made in building large knowledge bases, they are intrinsically incomplete. It is difficult to combine multiple bases due their different choices of representation and inference mechanisms, thereby limiting users one base its reasonable methods for any specific task. This paper presents a multi-agent framework integration, proposes an approach capability modeling without common ontology. The proposed model provides general description heterogeneous such that contents accessible by the knowledge-based agents may be matched up against requests. concept correlation matrix transformed into k-dimensional vector space using low-rank approximation dimensionality reduction. Experiments performed with matchmaking mechanism integration models ConceptNet, WordNet, Wikipedia. In user study, results compared ranked lists produced online show over 85% them accurate have positive user-produced lists.