Towards Web-Based Semantic Knowledge Completion for Adaptive World Modeling in Cognitive Systems

摘要

This contribution proposes web-based concept completion, an approach for enriching concept descriptions created from observations with additional semantic information. The proposed methods are designed as an extension to the Object-Oriented World Model (OOWM), a system for environment representation in cognitive systems using semantic knowledge captured in a domain model. An adaptive version of the OOWM allows extending the domain model with new concept descriptions created from observations. Such sensor observations yet only provide the perceptual information part of a concept description and lack semantic information. To this end, this contribution is concerned with selecting and integrating suitable resources readily available from the World Wide Web into a demonstrator for concept completion.