Skip to main content

> ML_ALGORITHM // RELATIONAL-GRAPH-CONVOLUTIONAL-NETWORKS-RGCN_v1.0

Relational Graph Convolutional Networks (R-GCN)

Extension of GCNs designed for highly multi-relational directed graphs and knowledge bases using relation-specific transformation weights.

Heterogeneous Graph Neural Networksgraph-relationalmoderate-posthoclarge (>100k)
Back to All Algorithms
Computational Complexity
Training Complexity:O(|E| * d + relations * bases * d^2)
Inference Complexity:O(|E| * d)
Hardware Profile
CPU Friendly:No
Requires GPU:Yes
Memory Footprint:high
Interpretability & Data
Interpretability Tier:moderate-posthoc
Training Data Needs:large (>100k)

Interpretability Assessment

Basis sharing reveals structural similarities across different multi-relational edge types.

Suitable Tasks & Supported Modalities

Suitable Tasks:
knowledge graph completionheterogeneous node classification
Supported Modalities:
graph

Implementing Libraries

torch-geometric
Deep Graph Library (DGL)DMLC / AWS / NYU · v2.2.1
View Spec

Foundational Literature

Common Pitfalls & Warnings
  • Parameter explosion on graphs with thousands of relation types without regularized basis decomposition