# Bi-Level Graph Neural Networks for Drug-Drug Interaction Prediction

arXiv: https://arxiv.org/abs/2006.14002 · HTML: https://mlsanity.com/paper/2006.14002/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Jun 2020.

## Built on by (most cited first, 8 of 8)

- [Graph Neural Networks for Molecules](https://mlsanity.com/paper/2209.05582/) (2022)
- [HyGNN: Drug-Drug Interaction Prediction via Hypergraph Neural Network](https://mlsanity.com/paper/2206.12747/) (2022)
- [LEGO-Learn: Label-Efficient Graph Open-Set Learning](https://mlsanity.com/paper/2410.16386/) (2024)
- [A Unified View of Relational Deep Learning for Drug Pair Scoring](https://mlsanity.com/paper/2111.02916/) (2021)
- [Hierarchical Graph Representation Learning for the Prediction of Drug-Target Binding Affinity](https://mlsanity.com/paper/2203.11458/) (2022)
- [Graph-structured Small Molecule Drug Discovery Through Deep Learning: Progress, Challenges, and Opportunities](https://mlsanity.com/paper/2502.08975/) (2025)
- [Recent Developments in GNNs for Drug Discovery](https://mlsanity.com/paper/2506.01302/) (2025)
- [Zero-Shot Learning with Subsequence Reordering Pretraining for Compound-Protein Interaction](https://mlsanity.com/paper/2507.20925/) (2025)

## Abstract (arXiv)

We introduce Bi-GNN for modeling biological link prediction tasks such as drug-drug interaction (DDI) and protein-protein interaction (PPI). Taking drug-drug interaction as an example, existing methods using machine learning either only utilize the link structure between drugs without using the graph representation of each drug molecule, or only leverage the individual drug compound structures without using graph structure for the higher-level DDI graph. The key idea of our method is to fundamentally view the data as a bi-level graph, where the highest level graph represents the interaction between biological entities (interaction graph), and each biological entity itself is further expanded to its intrinsic graph representation (representation graphs), where the graph is either flat like a drug compound or hierarchical like a protein with amino acid level graph, secondary structure, tertiary structure, etc. Our model not only allows the usage of information from both the high-level interaction graph and the low-level representation graphs, but also offers a baseline for future research opportunities to address the bi-level nature of the data.