Conference
ECAI 2026 — 18th Edition International Conference on Electronics, Computers and Artificial Intelligence
Bucharest, Romania (online presentation)
Presented Hierarchical Hypergraph Neural Networks for Multi-Type Drug–Drug Interaction Prediction. The talk introduced the two-stage architecture behind my master's research: a chemical hypergraph learned from SMILES substructures, feeding a second network that classifies which of 86 interaction types occurs. The work argues that detecting an interaction and identifying its type are not separate problems but a single nested one — trained entirely on CPU.
