Show simple item record

dc.contributor.authorKong, Kezhi
dc.contributor.authorLi, Guohao
dc.contributor.authorDing, Mucong
dc.contributor.authorWu, Zuxuan
dc.contributor.authorZhu, Chen
dc.contributor.authorGhanem, Bernard
dc.contributor.authorTaylor, Gavin
dc.contributor.authorGoldstein, Tom
dc.date.accessioned2020-11-04T11:06:30Z
dc.date.available2020-11-04T11:06:30Z
dc.date.issued2020-10-19
dc.identifier.urihttp://hdl.handle.net/10754/665805
dc.description.abstractData augmentation helps neural networks generalize better, but it remains an open question how to effectively augment graph data to enhance the performance of GNNs (Graph Neural Networks). While most existing graph regularizers focus on augmenting graph topological structures by adding/removing edges, we offer a novel direction to augment in the input node feature space for better performance. We propose a simple but effective solution, FLAG (Free Large-scale Adversarial Augmentation on Graphs), which iteratively augments node features with gradient-based adversarial perturbations during training, and boosts performance at test time. Empirically, FLAG can be easily implemented with a dozen lines of code and is flexible enough to function with any GNN backbone, on a wide variety of large-scale datasets, and in both transductive and inductive settings. Without modifying a model's architecture or training setup, FLAG yields a consistent and salient performance boost across both node and graph classification tasks. Using FLAG, we reach state-of-the-art performance on the large-scale ogbg-molpcba, ogbg-ppa, and ogbg-code datasets.
dc.publisherarXiv
dc.relation.urlhttps://arxiv.org/pdf/2010.09891
dc.rightsArchived with thanks to arXiv
dc.titleFLAG: Adversarial Data Augmentation for Graph Neural Networks
dc.typePreprint
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
dc.contributor.departmentElectrical Engineering Program
dc.contributor.departmentVCC Analytics Research Group
dc.eprint.versionPre-print
dc.contributor.institutionUniversity of Maryland.
dc.contributor.institutionUS Naval Academy.
dc.identifier.arxivid2010.09891
kaust.personLi, Guohao
kaust.personGhanem, Bernard
refterms.dateFOA2020-11-04T11:07:27Z


Files in this item

Thumbnail
Name:
Preprintfile1.pdf
Size:
426.2Kb
Format:
PDF
Description:
Pre-print

This item appears in the following Collection(s)

Show simple item record