Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Hypergraph Pattern Machine: Compositional Tokenization for Higher-Order Interactions

About

Hypergraphs model higher-order relations that drive real-world decisions, from drug prescriptions to recommendations. A central structural signal in such data, beyond what pairwise relations can express, is interaction compositionality: whether a higher-order relation is compositional, emergent, or inhibitory with respect to its observed or unobserved sets. In polypharmacy, the regime decides whether a drug should be dropped, kept, or excluded: a compositional drug triple can be safely simplified, an emergent triple requires all drugs jointly, and an inhibitory triple flags a drug that disrupts an existing interaction. However, existing hypergraph learning methods, which merely propagate messages over observed hyperedges, leave this compositional signal unmodeled, allowing dangerous drug combinations to slip through and be misclassified. To this end, we propose the Hypergraph Pattern Machine (HGPM), shifting the paradigm from message passing to learning the compositional pattern of subsets. It tokenizes compositional subsets, organizes them in an inclusion DAG, and trains an inclusion-aware Transformer under masked reconstruction. On ten hypergraph benchmarks, HGPM matches or exceeds state-of-the-art methods. Notably, in a real adverse-event prediction case, HGPM correctly identifies the drug addition that inhibits the side effect among feature-identical candidates, a discrimination existing methods cannot make. The code and data are in https://github.com/KryieZhao/HGPM.git.

Kyrie Zhao, Zehong Wang, Tianyi Ma, Fang Wu, Xiangru Tang, Pietro Lio, Sheng Wang, Yanfang Ye• 2026

Related benchmarks

TaskDatasetResultRank
Hypergraph Node ClassificationCiteseer 50/25/25 (test)
Test Accuracy77.1
16
Hypergraph Node ClassificationPubmed 50/25/25 (test)
Test Accuracy89.7
16
Hypergraph Node ClassificationHouse 50/25/25 (test)
Test Accuracy79.2
16
Hypergraph Node ClassificationDBLP-CA 50/25/25 (test)
Test Accuracy92.2
15
Hypergraph Node ClassificationSenate 50/25/25 (test)
Test Accuracy77.2
14
Hypergraph Node ClassificationWalmart 50/25/25 (test)
Test Accuracy71.2
14
Hypergraph Node ClassificationCora-CA 50/25/25 (test)
Test Accuracy85.1
14
Hypergraph Node ClassificationCongress 50/25/25 (test)
Test Accuracy94.1
13
Edge classificationHODDI (test)
F1 Score92.9
8
Edge classificationJADER (test)
F1 Score78.4
8
Showing 10 of 12 rows

Other info

Follow for update