Ontology String Mapping via Submodular Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for mapping sets of strings onto an ontology face challenges, particularly when relationships between the strings are unclear, as they often rely on symmetric assumptions and similarity propagation techniques that are not feasible in cases where the notion of 'neighbor' is lacking, leading to missed distant meanings and reduced ambiguity.
Innovation Solution
A method that computes pairwise similarities between strings and ontology objects, calculates global scores using a submodular function to identify an optimal subset, and maps strings based on these similarities, with the option of assigning a null node if no suitable match is found, effectively balancing local and global parameters to achieve coherent mapping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If similarity propagation techniques are used for mapping strings onto ontology, then local matching accuracy is improved, but the method fails when the notion of 'neighbor' is lacking in the input set of strings
Solution Approach 1:
The patent transforms the symmetric similarity propagation assumption into an asymmetric framework by introducing a global submodular function that evaluates candidate subsets independently. Instead of assuming symmetry in neighbor relationships, the method asymmetrically handles cases where neighbor notions are lacking by directly computing global scores based on submodular optimization, thus adapting to asymmetric or undefined relationship structures in the input data.
Solution Approach 2:
The patent changes the evaluation parameter from local pairwise similarity propagation to global submodular function optimization. By transitioning from incremental similarity propagation to direct global scoring using submodular functions, the method adapts to cases where traditional neighbor-based parameters are not applicable, maintaining matching accuracy without requiring defined neighbor relationships.
2Stability of the object's composition
If dense subgraph method is used to reduce ambiguity, then interpretation coherence is improved, but distant meanings are missed
Solution Approach 1:
The patent moves from a local two-dimensional view (pairwise matching and dense subgraph connectivity) to a global three-dimensional evaluation space that incorporates submodular function scores. This dimensional expansion allows the method to evaluate candidate subsets based on global coherence metrics rather than just local density, thereby capturing distant meanings that would be missed in purely local dense subgraph approaches while maintaining interpretation coherence.
Solution Approach 2:
The patent creates a universal evaluation framework using submodular functions that can handle both local coherent interpretations and distant relationships simultaneously. The submodular function serves multiple purposes: evaluating local consistency, capturing global coherence, and identifying valid candidate subsets regardless of their positional relationships, thus preventing loss of distant meanings while maintaining interpretation stability.
3Reliability
If global matching techniques are used to construct optimal alignment, then mutual impact of pairwise decisions is considered, but computational complexity increases
Solution Approach 1:
The patent replaces the mechanical iterative similarity propagation process with a direct submodular function optimization approach. Instead of mechanically propagating similarities through iterative neighbor comparisons, the method substitutes this with a mathematical optimization framework based on submodular functions, which can directly compute global scores for candidate subsets. This substitution reduces computational complexity while maintaining the reliability of coherent mapping by capturing mutual impacts of pairwise decisions in the global optimization objective.
Data Source
AI summary
A computer-implemented system and method provide for mapping a set of strings onto an ontology which may be represented as a graph. The method includes receiving a set of strings, each string denoting a respective object. For each of the strings, a pairwise similarity is computed between the string and each of a set of objects in the ontology. For each of a set of candidate subsets (subgraphs) of the set of objects, a global score is computed, which is a function of the pairwise similarities between the strings and the objects in the subset and a tightness score. The tightness score is computed on the objects in the subset with a submodular function. An optimal subset is identified from the set of candidate subsets based on the global scores. Strings in the set of strings are mapped to the objects in the optimal subset, based on the pairwise similarities.