A Survey of Knowledge Graph Reasoning on Graph Types: Static, Dynamic, and Multi-Modal.

Area of Science:

• Artificial Intelligence
• Data Science
• Computer Science

Background:

• Knowledge Graph Reasoning (KGR) deduces new facts from existing data using logic rules.
• KGR enhances AI applications like question answering and recommendation systems.
• Existing KGR models are categorized into static, temporal, and multi-modal types.

Purpose of the Study:

• To provide the first comprehensive survey of Knowledge Graph Reasoning (KGR) models.
• To systematically review KGR models from static to temporal and multi-modal approaches.
• To address the lack of comprehensive summaries and open-source repositories in KGR.

Main Methods:

• Models reviewed based on a bi-level taxonomy: graph types (static, temporal, multi-modal) and base-level (techniques, scenarios).
• Summarization of model performances and associated datasets.
• Identification of current challenges and future research opportunities in KGR.

Main Results:

• Categorization of KGR models into static, temporal, and multi-modal.
• Detailed review of techniques and scenarios within each category.
• Presentation of performance benchmarks and datasets for comparative analysis.

Conclusions:

• The survey offers a structured overview of the KGR landscape, bridging static, temporal, and multi-modal paradigms.
• It serves as a valuable resource for researchers by summarizing existing work, datasets, and performance metrics.
• The study highlights key challenges and opportunities, paving the way for advancements in knowledge graph reasoning.