Blog Posts
Short research notes, tutorials on Graph Machine Learning, Graph Neural Networks, missing features, and link prediction.
-
When Most Methods Look Robust: Rethinking GNN Benchmarks for Missing Node Features
Why current Graph Neural Network benchmarks for missing features are misleading, and a simple missing-indicator baseline (GNNmim) that holds up under realistic missingness.
Read the post →April 27, 2026 · 13 min read -
MCAR, MAR, and MNAR: A Practical Guide to Missing Data Mechanisms (with Python)
What MCAR, MAR, and MNAR actually mean, with three real-world scenarios — IoT sensors, demographic surveys, and clinical data — simulated in Python. Why the distinction changes which imputation methods are valid.
Read the post →April 26, 2026 · 15 min read