Co-reference Resolution via Domain Ontology Linking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current natural language processing (NLP) systems face challenges in resolving nontrivial semantic co-references in unstructured text data, particularly noun phrase co-references, which are essential for extracting meaningful information from vast digital sources in competitive environments like business intelligence, security, and counter-terrorism.
Innovation Solution
The method employs a domain knowledge ontology, such as a social network, to identify and link entities by using entity properties and relationships, leveraging a cognitive system that learns ontology inferences and lexicons to spot mentions and relationships, thereby resolving semantic co-references through entity and relationship linking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If domain knowledge ontology is used to resolve semantic co-references, then measurement precision of entity identification is improved, but device complexity increases
Solution Approach 1:
The patent introduces a domain knowledge ontology as an intermediary structure that mediates between unstructured text data and entity identification processes. The ontology serves as a knowledge base that stores predefined entities, relationships, and domain-specific information, enabling the system to resolve semantic co-references by matching text mentions against the ontology rather than relying on complex algorithmic reasoning alone.
Solution Approach 2:
The system performs preliminary action by pre-building and storing domain knowledge in the ontology before processing text data. Entities, relationships, and domain concepts are organized in advance in the ontology structure, allowing the co-reference resolution process to efficiently query and match against pre-processed knowledge rather than performing complex analysis during text processing.
2Measurement precision
If cognitive system learns ontology inferences and lexicons to spot mentions and relationships, then information extraction accuracy is improved, but loss of time in processing increases
Solution Approach 1:
The cognitive system performs preliminary learning of ontology inferences and lexicons before actual text processing. The system pre-processes the domain ontology to extract inference rules, relationship patterns, and lexical information, storing these in optimized structures that can be quickly applied during text analysis without repeating the full learning process.
Solution Approach 2:
The system creates simplified copies or representations of the learned knowledge in the form of lexicons and inference rules that can be quickly applied during text processing. Instead of performing full ontology learning during each text analysis, the system uses pre-compiled lexical resources and relationship patterns that capture the essential learned information in a more efficient format.
Data Source
AI summary
Embodiments for co-reference resolution and entity linking from unstructured text data by a processor. Semantic co-references and mentions of one or more entities may be resolved occurring in unstructured text data by linking the one or more entities using a domain knowledge ontology.