Ontology Extraction for Large Ontologies via Modularity and Forgetting

Abstract

We are interested in the computation of ontology extracts based on forgetting from large ontologies in real-world scenarios. Such scenarios require nearly all of the terms in the ontology to be forgotten, which poses a significant challenge to forgetting tools. In this paper we show that modularization and forgetting can be combined beneficially in order to compute ontology extracts. While a module is a subset of axioms of a given ontology, the solution of forgetting (also known as a uniform interpolant) is a compact representation of the ontology limited to a subset of the signature. The approach introduced in this paper uses an iterative workflow of four stages: (i) extension of the given signature and, if needed partitioning, (ii) modularization, (iii) forgetting, and (iv) evaluation by domain expert. For modularization we use three kinds of modules: localitybased, semantic and minimal subsumption modules. For forgetting three tools are used: Nui, Lethe and Fame. An evaluation on the SNOMED CT and NCIt ontologies for standard concept name lists showed that precomputing ontology modules reduces the number of terms that need to be forgotten. An advantage of the presented approach is high precision of the computed ontology extracts.

Ontologies and Data Management: A Brief Survey

“…UIs have been studied widely for DLs of various expressivity [243,244,277,284,301,371,383]. The same holds for the dual notion of forgetting, which is important for applications that require information hiding [123,150,237,382,389,390]; see also a recent survey in this journal [151].…”

Automated conceptual model clustering: a relator-centric approach

“…Techniques for model clustering, relevance, and summarization have been developed for more than three decades in conceptual modeling, from earlier approaches such as [16] in the mid-1980s, to [1,12] in the second half of the 1990s, to [47,48] in the first decade of the 2000s, to recent approaches such as [11]. Recently, the problem of complexity management of large symbolic models has gained significant interest also in the areas of ontology engineering [2,6,13,14,20,33], enterprise models [34], and process models [53,55].…”

“…Most of the methods for complexity management of large ontologies, however, consider these models simply as logical theories and are oblivious to these distinctions about generality level. These methods are divided in approaches for model abstraction (known as forgetting [13]), clustering (called ontology partitioning [6,20]), as well as ontology module extraction [13]. A few approaches such as [13] propose to combine operations of module extraction and forgetting.…”

Signature-Based Abduction for Expressive Description Logics

“…2. Existing methods for uniform interpolation perform best when the signature or its complement relative to the signature of the ontology are small [Koopmann and Schmidt, 2015;Zhao and Schmidt, 2017;Chen et al, 2019]. For signature-based abduction, we exploit the fact that the observation Ψ is usually small compared to K. Our method focuses on inferences relevant to ¬Ψ using a modified setof-support strategy.…”

“…Uniform interpolation for ALC is challenging, as solutions can in the worst case be of size triple-exponential in the size of the input [Lutz and Wolter, 2011]. This challenge does impact practical implementations, which usually only perform well with signatures that are either very small or very large [Koopmann and Schmidt, 2015;Zhao and Schmidt, 2017;Chen et al, 2019]. Moreover, as we would use K∧¬Ψ as input, most of the computed consequences would only depend on K and thus have no relevance to the abduction problem of explaining Ψ.…”