Confidential — Stefan Michaelcheck Only

A Simplex Approach to Synthetic Knowledge Graph Generation

2026benchmark creationnovelmethod

Ana Alexandra Morim da Silva, Axel-Cyrille Ngonga Ngomo, Atul Bhopalsing Pundir, Michael Röder

https://doi.org/10.1145/3774904.3792566OpenAlex: W7152614770
2
URLs Found
0
Internal Citations
4
Authors
usable
Abstract Quality
GPT-5.5 Abstract Analysis

Problems Identified (4)

Scalable KG Benchmarking: The growing scale of knowledge graphs requires scalable processing systems and large knowledge graphs for accurate benchmarking.

Missing Higher-Order Structures: Current data-driven synthetic knowledge graph generators operate at the triple level and do not consider higher-order graph structures.

Scalable KG Benchmarking: The growing scale of knowledge graphs requires scalable processing systems and large knowledge graphs for accurate benchmarking.

Missing Higher-Order Structures: Current data-driven synthetic knowledge graph generators operate at the triple level and do not consider higher-order graph structures.

Proposed Solutions (4)

SimplexKG: SimplexKG is a simplex-based synthetic knowledge graph generator that analyzes d-dimensional simplices in input knowledge graphs and uses the resulting simplicial networks to generate synthetic graphs of arbitrary size.

Higher-Dimensional Structure Modeling: The approach leverages higher-dimensional structures to improve the realism of generated synthetic knowledge graphs.

SimplexKG: SimplexKG is a simplex-based synthetic knowledge graph generator that analyzes d-dimensional simplices in input knowledge graphs and uses the resulting simplicial networks to generate synthetic graphs of arbitrary size.

Higher-Dimensional Structure Modeling: The approach leverages higher-dimensional structures to improve the realism of generated synthetic knowledge graphs.

Results (3)

Improved Structural Fidelity:

Improved Triple Store Benchmarking:

Improved Structural Fidelity:

Research Domain

synthetic knowledge graph generation and benchmarking

← Back to all papers