Reorganizing Ontologies: A Semantic Approach to Preserving Structure and Meaning

Ontologies are complex graphs that represent knowledge in a domain, while semantic subgraphs are smaller graphs extracted from these larger graphs to better understand their meaning. In this article, we explore the process of extracting semantic subgraphs from ontologies using a circuit model. The model adds "voltage" (information capacity) and "current" (semantic information flow) to each vertex in the graph, enabling us to rank triples and extract semantically meaningful subgraphs.
The algorithm for extracting semantic subgraphs involves adding one volt to each concept, solving a circuit equation, and then using greedy algorithms to find the k-size subgraph with maximum captured flow. This process results in a semantic subgraph that precisely describes the meaning of the concept or property. The authors propose organizing relevant literal information based on these semantic subgraphs as virtual documents called Semantic Description Documents (SDD).
The article demystifies complex concepts by using everyday language and engaging metaphors to explain complex ideas, such as "information capacity" and "semantic information flow." For example, the authors describe voltage as the ability of a circuit to convey information, while current represents how well that information is preserved. Similarly, they explain greedy algorithms as follows: "Imagine you are trying to pick the best candies from a bag. A greedy algorithm would take one candy at a time and always choose the biggest or most delicious one."
By focusing on demystifying complex concepts and using everyday language, the article achieves a balance between simplicity and thoroughness, providing readers with a comprehensive understanding of ontologies and semantic subgraphs without oversimplifying the material.