Graph Property Prediction on Open Graph Benchmark: A Winning Solution by Graph Neural Architecture Search
This work addresses graph classification tasks in computational chemistry and biology, offering an incremental improvement through neural architecture search.
The authors tackled graph property prediction on molecular and protein datasets from the Open Graph Benchmark by developing a graph neural network framework that incorporates pooling architecture search and feature selection strategies, achieving a performance breakthrough significantly better than fixed-aggregate methods.
Aiming at two molecular graph datasets and one protein association subgraph dataset in OGB graph classification task, we design a graph neural network framework for graph classification task by introducing PAS(Pooling Architecture Search). At the same time, we improve it based on the GNN topology design method F2GNN to further design the feature selection and fusion strategies, so as to further improve the performance of the model in the graph property prediction task while overcoming the over smoothing problem of deep GNN training. Finally, a performance breakthrough is achieved on these three datasets, which is significantly better than other methods with fixed aggregate function. It is proved that the NAS method has high generalization ability for multiple tasks and the advantage of our method in processing graph property prediction tasks.