GABO: Graph Augmentations with Bi-level Optimization

Abstract

Data augmentation refers to a wide range of techniques for improving model generalization by augmenting training examples. Oftentimes such methods require domain knowledge about the dataset at hand, spawning a plethora of recent literature surrounding automated techniques for data augmentation. In this work we apply one such method, bilevel optimization, to tackle the problem of graph classification on the ogbg-molhiv dataset. Our best performing augmentation achieved a test ROCAUC score of 77.77 % with a GIN+virtual classifier, which makes it the most effective augmenter for this classifier on the leaderboard. This framework combines a GIN layer augmentation generator with a bias transformation and outperforms the same classifier augmented using the state-of-the-art FLAG augmentation.

References in corpus (5)

• The Marginal Value of Adaptive Gradient Methods in Machine Learning
• Open Graph Benchmark: Datasets for Machine Learning on Graphs
• Learning Graph-Level Representation for Drug Discovery
• Towards Theoretically Understanding Why SGD Generalizes Better Than ADAM in Deep Learning
• MoleculeNet: A Benchmark for Molecular Machine Learning