Repairing Networks of $\mathcal{EL_\perp}$ Ontologies using Weakening and Completing -- Extended version

TL;DR

This work addresses repairing networks of ${\mathcal{EL_\perp}}$ ontologies by avoiding the excessive removal of correct domain knowledge when removing wrong axioms. It extends prior single-ontology repair frameworks with formal problem definitions for ontology networks, introduces multiple autonomy levels and combination operators, and analyzes how these choices affect repair quality and the burden of domain- expert validation. The authors formalize DP repairs, explore background-knowledge and add-set configurations, and demonstrate that operator choices materially impact completeness and correctness, supported by experiments on multiple OAEI-derived networks. An implemented Java-based tool extends the EL RepOSE system to support debugging, weakening, and completing in network repairs, offering interactive, visual guidance and a blueprint for extending existing alignment-repair systems to general ontology networks.

Abstract

The quality of ontologies and their alignments is crucial for developing high-quality semantics-based applications. Traditional debugging techniques repair ontology networks by removing unwanted axioms and mappings, but may thereby remove consequences that are correct in the domain of the ontology network. In this paper we propose a framework for repairing ontology networks that deals with this issue. It defines basic operations such as debugging, weakening and completing. Further, it defines combination operators that reflect choices in how and when to use the basic operators, as well as choices regarding the autonomy level of the ontologies and alignments in the ontology network. We show the influence of the combination operators on the quality of the repaired network and present an implemented tool. By using our framework together with existing algorithms for debugging, weakening and completing, we essentially provide a blueprint for extending previous work and systems.

Figures (6)

• Figure 1: Debugging, weakening and completing.
• Figure 2: Hasse diagrams (a from LL24LL24submitted and b-d from LL23). (a) selecting and debugging; (b) removing and adding back wrong axioms; (c) weakening and updating; (d) completing and updating.
• Figure 3: The network contains unsatisfiable concepts E, F, e, and f. Asserted axiom e $\sqsubseteq$ b in $O_{2}$, and mapping b $\sqsubseteq$ D between $O_{2}$ and $O_{1}$ lead to network incoherence.
• Figure 4: Hasse diagrams regarding (a) the use of different knowledge bases during repairing (KB); (b) different restrictions on the add set (AS);
• Figure 5: Sub and sup when weakening wrong axiom Organizator $\sqsubseteq$ Organizing_Committee_Member in ontology network ekaw-sigkdd using (a) KB$_O$; (b) KB$_{ON}$. Concepts in wrong axiom in red, concepts in ekaw in blue, concepts in sigkdd in black.

Theorems & Definitions (4)

• Definition 1
• Definition 2
• Definition 3
• Definition 4