FSDiffNet is a package for fast and convenient differential graph inference based on siamese neural network.

Analysis of temporal fluctuations in Stack Overflow user activity to characterize interaction-network structure, test fluctuation-based null models, and infer candidate connections through lagged covariance and inverse-network methods, developed for HUJI’s Dynamics, Networks and Computation course (67562).

GSM artifact for exact online GNN inference serving with degree-based scheduling and memory-aware batch division.

AMIE+ rule mining and inference engine for Neo4j knowledge graphs — discover logical rules, then infer new relationships automatically.