English

Elastic Weight Consolidation for Knowledge Graph Continual Learning: An Empirical Evaluation

Machine Learning 2025-12-02 v1

Abstract

Knowledge graphs (KGs) require continual updates as new information emerges, but neural embedding models suffer from catastrophic forgetting when learning new tasks sequentially. We evaluate Elastic Weight Consolidation (EWC), a regularization-based continual learning method, on KG link prediction using TransE embeddings on FB15k-237. Across multiple experiments with five random seeds, we find that EWC reduces catastrophic forgetting from 12.62% to 6.85%, a 45.7% reduction compared to naive sequential training. We observe that the task partitioning strategy affects the magnitude of forgetting: relation-based partitioning (grouping triples by relation type) exhibits 9.8 percentage points higher forgetting than randomly partitioned tasks (12.62% vs 2.81%), suggesting that task construction influences evaluation outcomes. While focused on a single embedding model and dataset, our results demonstrate that EWC effectively mitigates catastrophic forgetting in KG continual learning and highlight the importance of evaluation protocol design.

Keywords

Cite

@article{arxiv.2512.01890,
  title  = {Elastic Weight Consolidation for Knowledge Graph Continual Learning: An Empirical Evaluation},
  author = {Gaganpreet Jhajj and Fuhua Lin},
  journal= {arXiv preprint arXiv:2512.01890},
  year   = {2025}
}

Comments

Accepted to NORA Workshop at NeurIPS 2025

R2 v1 2026-07-01T08:04:07.629Z