SKMT: A Unified Ontological Knowledge Platform for Semantic Content Tagging and Search

Dr. Jiangbo Dang
Siemens Corporate Research


It has become increasingly difficult to share knowledge or locate the right information and people across various organization resources. Corporations are always interested in managing and sharing intellectual assets, and maintaining an exponentially increasing number of content. Nowadays we heavily rely on search engines to locate information. However, existing technologies are experiencing difficulties: keyword based search often return results with low precision and recall. An approach for mitigating this issue is to use content tags. Content tagging helps users to describe and organize content. Good tags provide relevant and brief information about resources. Such tagging systems have major limitations because user generated and/or selected tags are (1) free from context and form, (2) biased, (3) used for purposes other than description, and (4) often ambiguous. Since tagging is a subjective, time-consuming voluntary work, most documents are not tagged at all. Semantic web technologies can be utilized to automatically generate semantic tags. Semantic tags not only reflect document content more accurately, they also enable better search results. Ontology coverage, ontology mapping and weighting significant ontological entities within a context are key challenges in semantic tagging and search systems. To address these challenges, we proposed a Semantic Knowledge Management (SKMT) platform that would leverage usage of intelligent semantic content tagging, indexing and search methods. First, we built a UNIpedia to cover most named English terms. UNIpedia is an ontological knowledge base that unifies different ontological knowledge bases by reconciling their instances as WordNet concepts. Secondly, we developed a semantic tagging system to map free text to semantic tags defined as entities in an ontology. Semantic TagPrint uses a linear time lexical chaining Word Sense Disambiguation (WSD) algorithm for real time concept mapping. Finally we implemented a Semantic search tool for indexing and searching various data sources. It provides a user-friendly interface that allows users to manage knowledge and data sources, scan and index them, explore and visualize result data.