Reproduce “Time-aware Random Walk Diffusion to Improve Dynamic Graph Learning” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Jong-whi Lee, Jinhong Jung

TIARA augments discrete-time graph sequences for dynamic graph neural networks. It uses random walks combining current-snapshot connections with proximity carried forward from earlier snapshots. This adds spatial and temporal locality before model training. On BitcoinAlpha, TIARA-augmented GCRN scores 94.6 ± 0.8 AUC, versus 93.1 ± 0.4 for MERGE; however, Reddit results remain below random for GCN and EvolveGCN.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.