ABox Abstraction Refinement for Scalable Ontology Reasoning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for ontology reasoning, such as query rewriting and full materialization, face challenges with increasing data volumes due to limitations in metadata expressivity, performance, storage space, and memory usage, particularly in handling large ABoxes.

Innovation Solution

The method employs ABox abstraction and refinement, which involves normalizing the TBox and computing a compressed ABox representation, allowing for efficient materialization of concepts without the need for resource-intensive justification computations, thereby reducing computing and memory requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If full materialization is used to pre-compute all entailments, then query answering performance is improved, but storage space and memory requirements increase significantly

Engineering Contradiction:
Improvequery answering performanceVSAvoidstorage space
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The patent segments the ABox into multiple partitions based on concept assertions, where each partition contains individuals sharing the same concept assertions. This segmentation allows the system to process and materialize entailments for smaller partitions independently, reducing peak memory requirements while maintaining complete query answering capability across all partitions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies partial materialization by computing entailments only for individuals that are relevant to the query at hand, rather than pre-computing all possible entailments for the entire ABox. The system materializes entailments incrementally as needed, balancing between query performance and resource consumption by avoiding unnecessary computation and storage of unrelated entailments.

Inventive Principle:
Principle #16Partial or excessive action

2Quantity of substance

If query rewriting is used to handle metadata, then storage requirements are reduced, but metadata expressivity is limited

Engineering Contradiction:
Improvestorage requirementsVSAvoidmetadata expressivity
Core Design Contradiction:
Quantity of substanceVSAdaptability or versatility

Solution Approach 1:

The patent introduces an intermediary approach by using a hybrid architecture that combines elements of both query rewriting and materialization. The system uses concept materialization for individuals within partitions while maintaining the ability to handle expressive OWL 2 DL metadata through the underlying reasoning engine, thus achieving both storage efficiency and metadata expressivity.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If combined approaches are used to materialize some entailments in advance, then query performance is improved, but the approach needs to be adjusted and tuned for each application case

Engineering Contradiction:
Improvequery performanceVSAvoidconfiguration complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements self-service by automatically partitioning the ABox based on concept assertion patterns and selecting appropriate materialization strategies without requiring manual configuration. The system autonomously determines which individuals to materialize based on query patterns and data characteristics, eliminating the need for application-specific tuning while maintaining optimized query performance.

Inventive Principle:
Principle #25Self-service

4Quantity of substance

If ABox size increases to handle large data volumes, then data coverage is improved, but reasoning performance and memory usage deteriorate

Engineering Contradiction:
Improvedata coverageVSAvoidreasoning performance
Core Design Contradiction:
Quantity of substanceVSProductivity

Solution Approach 1:

The patent divides large ABoxes into multiple smaller partitions based on concept assertion equivalence, allowing the reasoning system to process each partition independently with reduced memory footprint. This segmentation enables the system to handle large data volumes by distributing the reasoning workload across multiple manageable units while maintaining overall reasoning performance.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transforms the single-dimensional problem of handling large ABoxes by introducing a new dimension of partitioning based on concept assertion patterns. Instead of processing individuals in a flat sequence, the system organizes them into hierarchical partitions that can be processed in parallel or sequentially based on query requirements, effectively adding a structural dimension that improves scalability.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP2966600B1Abstraction refinement for scalable type reasoning in ontology-based data repositories
Publication Date: 2018.08.15 DERIVO
  • EP2966600B1 patent drawingFigure 1
  • EP2966600B1 patent drawingFigure 2
  • EP2966600B1 patent drawingFigure 3

AI summary

A computer-implemented method for computing a concept materialization of an ontology is presented whereby a compression technique called "ABox abstraction and refinement" is used that may significantly reduce time, memory, and computing resources for reasoning and in particular for computing and outputting the materialization.