EdgeReMIND: A Scalable, Top-Ranked Memorization Baseline for Temporal Multi-Relational Link Prediction

arXiv:2609.17916 · cs.LG · Submitted 2026-09-15 · Read on arXiv

cs.LG

Submitted: 2026-09-15

Updated: 2026-09-15

Comments: 22 pages, 6 figures, 13 tables. Accepted at the Fifth Learning on Graphs Conference (LoG 2026), Proceedings Track. Code: https://github.com/BryantPollard/EdgeReMIND

Code: https://github.com/BryantPollard/EdgeReMIND

License: http://creativecommons.org/licenses/by/4.0/

The gist: Temporal link prediction on the Temporal Graph Benchmark 2.0 (TGB 2.0) faces a scalability ceiling: on the benchmark's three largest datasets, every existing embedding method runs out of memory or

Terminology

Abstract

Temporal link prediction on the Temporal Graph Benchmark 2.0 (TGB 2.0) faces a scalability ceiling: on the benchmark's three largest datasets, every existing embedding method runs out of memory or exceeds the time budget. These large-scale graphs are the ones nearest real deployment scale, so failing on them is a real production limitation. EdgeReMIND sets the highest reported test mean reciprocal rank (MRR) on six of eight TGB 2.0 datasets and is the only relation-aware method that runs on all of them. This linear memorization model, with learned per-relation weights over data-calibrated features, is therefore not merely a fallback where embeddings fail but a practical state-of-the-art baseline across the benchmark.

Sources

Related papers