Hypergraph Pattern Machine Models Higher-Order Drug Interactions

A new machine learning model, the Hypergraph Pattern Machine (HGPM), addresses a critical gap in hypergraph learning by modeling interaction compositionality in higher-order relations. Unlike existing methods that only propagate messages over observed hyperedges, HGPM distinguishes between compositional, emergent, and inhibitory interactions. In polypharmacy, this distinction determines whether a drug combination is safe to simplify, requires all drugs jointly, or should be excluded. The model aims to prevent dangerous drug combinations from being misclassified. The research is published on arXiv under identifier 2605.16527.