SEMANTIC BRIDGE: A MODEL FOR SEMANTIC WEB LITERATURE-BASED DISCOVERIES

Abstract

Literature-based discovery focuses on retrieving hypotheses from apparently unrelated scientific papers. Studies show that joining facts from different documents can suggest unpublished hypotheses, potentially leading to significant discoveries. With over fifty million papers published and the significant benefits to humanity that discoveries bring, succeeding in automatically generating new hypotheses from the scientific literature is both challenging and crucial. At this stage, a collective effort and a strong collaboration of technical and domain experts of different disciplines are required. The best system can be built with technologies that standardize knowledge sharing and reasoning by combining published data, like those offered by the Semantic Web mechanisms. Surprisingly, existing systems proposed for making discoveries overlook these tools and neither provide means to include other tools to resolve a specific part of the problem. Thus, they require considerable effort to build from the ground up, which confines them to merely interesting projects that are eventually abandoned. One reason for not adopting Semantic Web technologies could be the previously scarce availability of essential data resources, a situation that has changed rapidly.

Literature-based discovery systems based on Semantic Web technologies can now be built; the possibility of having a system to make discoveries where researchers, practitioners, and experts of different disciplines collaborate to create its building blocks is concrete.

With my thesis, I propose a model that uses Semantic Web technologies to facilitate the development of collaborative, expandable, and reusable discovery systems. A proof-of-concept will be presented, demonstrating the integration of natural language processing, efficient reasoning, and innovative result presentation methods. This system will highlight the effectiveness of publishing scientific assertions with nanopublications, preserving provenance and intellectual property, suggesting a new way to publish papers, and the emergence of a Web of scientific facts for which the study aims to lay the foundation.