A Review of Knowledge Graphs: Modern Frameworks, Applications, and Open Challenges
Kumkum Saxena, Gaurav Kukdeja, Arohi Jambenal, Piyush Hingorani
Lecture notes in networks and systems
Problems Identified (5)
KG construction and deployment complexity: Knowledge graph frameworks and deployment span manual curation, automatic extraction, and hybrid paradigms that require review and comparison.
KG scalability and interoperability challenges: Knowledge graphs face challenges including data sparsity, computational complexity, standardization, and multimodal data integration.
Compound SPARQL query difficulty: Even strong models struggle to handle compound SPARQL queries effectively over RDF-based knowledge graphs.
KG construction and deployment complexity: Knowledge graph frameworks and deployment span manual curation, automatic extraction, and hybrid paradigms that require review and comparison.
KG scalability and interoperability challenges: Knowledge graphs face challenges including data sparsity, computational complexity, standardization, and multimodal data integration.
Proposed Solutions (5)
KG frameworks and deployment review: The paper reviews knowledge graph frameworks and deployment methods across manual, automatic, and hybrid construction paradigms.
LLM-KG integration review: The paper investigates integration of knowledge graphs with large language models and discusses contributions to knowledge representation, entity resolution, and context-aware retrieval.
KG research agenda: The paper identifies future research directions such as multimodal knowledge fusion, query execution optimization, and self-adapting knowledge graphs.
Recent KG literature review: The paper provides a synopsis of knowledge graphs and reviews methodologies, frameworks, and recent achievements from the last three years.
KG frameworks and deployment review: The paper reviews knowledge graph frameworks and deployment methods across manual, automatic, and hybrid construction paradigms.
Results (3)
KG AI application coverage:
KG integration benefits:
LLM support and SPARQL limitation:
Research Domain
Knowledge graphs and AI applications