Human-Based Semantic Web Services: Phase 1

This paper describes an approach to address the DERI Stanford SWS 2006 challenge that explicitly considers human agents in the loop of both the data and process mediation, while aiming to minimize the involvements of such agents in subsequent usages. Our approach is based on the thesis that reaching full automation of data and process mediation is best achieved by evolving an approach that maintains humans in the loop and gradually refines the semantics from the human agents’ feedback and usage. This approach has also a nice side-effect that it can be trusted early on, since humans are involved in the process; this makes it more realistic to be implemented and likely to be deployed sooner in real world settings than one that tries to fully automate the mediations from the beginning.