Leveraging Classical Algorithms for Graph Neural Networks
This work addresses the challenge of enhancing GNN performance for molecular prediction, offering incremental improvements by embedding classical algorithmic priors.
The paper tackled the problem of improving Graph Neural Networks' (GNNs) generalization on molecular property prediction tasks by pretraining them on classical algorithms, resulting in absolute gains of 6% on ogbg-molhiv and 3% on ogbg-molclintox compared to a baseline.
Neural networks excel at processing unstructured data but often fail to generalise out-of-distribution, whereas classical algorithms guarantee correctness but lack flexibility. We explore whether pretraining Graph Neural Networks (GNNs) on classical algorithms can improve their performance on molecular property prediction tasks from the Open Graph Benchmark: ogbg-molhiv (HIV inhibition) and ogbg-molclintox (clinical toxicity). GNNs trained on 24 classical algorithms from the CLRS Algorithmic Reasoning Benchmark are used to initialise and freeze selected layers of a second GNN for molecular prediction. Compared to a randomly initialised baseline, the pretrained models achieve consistent wins or ties, with the Segments Intersect algorithm pretraining yielding a 6% absolute gain on ogbg-molhiv and Dijkstra pretraining achieving a 3% gain on ogbg-molclintox. These results demonstrate embedding classical algorithmic priors into GNNs provides useful inductive biases, boosting performance on complex, real-world graph data.