580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How can ontologies support automated composition and discovery of Knowledge Discovery in Databases (KDD) algorithms and processes?
This research theme focuses on the use of ontologies to model KDD algorithms, their properties, functionalities, and interrelations, enabling automatic or semi-automatic composition of KDD workflows to support users in selecting and assembling suitable algorithms for specific knowledge discovery tasks. This matters because KDD involves many complex tools and processes, and ontology-driven techniques can reduce the cognitive load on users by providing meaningful semantic abstractions and goal-driven composition strategies.
Key finding: Introduced a goal-driven procedure leveraging the KDDONTO ontology to automatically compose valid KDD algorithm workflows. The ontology formally models KDD algorithms at an abstract level, facilitating the generation and... Read more
Key finding: Presented KDDONTO as a formal ontology describing KDD algorithms to aid discovery and process composition. The paper emphasized ontology development methodologies adapted for KDD, formalizing domain concepts and relations to... Read more
Key finding: Developed a service-oriented architecture where KDD tools are described semantically at multiple abstraction levels - algorithm, tool, and service. This architecture uses ontologies to represent tool syntax and semantics... Read more
Key finding: Discussed ontological engineering principles relevant to building reusable ontologies across domains, underlying the formalization of KDDONTO-like ontologies. Presented methodologies, principles, and tools that support... Read more
2. What are the challenges and solutions in achieving semantic interoperability among diverse upper ontologies for consistent knowledge integration?
This theme investigates methods for relating multiple existing upper-level ontologies to enable semantic interoperability—critical for integrating knowledge across heterogeneous domains and systems. Upper ontologies provide foundational conceptual categories that domain ontologies build upon; bridging these is essential for reusing knowledge and aligning semantic interpretations, which impacts KDD and broader knowledge engineering applications.
Key finding: Reported a collaborative initiative to promote relations among upper ontologies by proposing mechanisms such as creating a Compatible Subset Upper Ontology (CSUO), merging, or alignment to facilitate semantic... Read more
Key finding: Developed GOL as an expressive upper-level ontology extending beyond set-theoretic foundations with ontologically basic categories including sets, individuals, and universals, as well as refined classifications. This work... Read more
Key finding: Presented BFO, a widely adopted top-level ontology designed to support integration across scientific domains, emphasizing principles like ontological realism and fallibilism. Demonstrated how BFO can represent change and... Read more
3. How can ontology engineering methodologies and design patterns improve ontology development and usability in domain-specific applications?
This theme explores methodologies, design patterns, and engineering practices that enhance the construction, maintenance, and domain adequacy of ontologies. Given the challenges non-expert users face in ontology building, especially in business and KDD contexts, developing reusable patterns and methodological frameworks supports ontology accuracy, expressiveness, and end-user friendliness—crucial for ontology adoption and effective semantic applications.
Key finding: Proposed OPAL, a set of high-level domain-specific ontology design patterns tailored for business experts to facilitate ontology building without requiring deep knowledge engineering expertise. By formalizing these design... Read more
Key finding: Provided an extensive overview of ontology engineering as a maturing discipline, discussing its evolution from knowledge management to an engineering practice integrated with software and business architectures. The work... Read more
Key finding: Presented the EC-DOC scheme that integrates domain analysis, conceptualization, fuzzy clustering, and localization to construct collaborative, reusable, and structured domain ontologies. By focusing on ecological and confined... Read more
Key finding: Discussed the bidirectional method for developing ontologies combining top-down conceptual modeling with bottom-up automatic extraction from linguistic corpora, emphasizing the integration of domain knowledge and... Read more