Adversarial Graph Augmentation to Improve Graph Contrastive Learning
Dec 6, 2021
Speakers
Susheel Suresh
Sprecher:in · 0 Follower:innen
Jennifer Neville
Sprecher:in · 0 Follower:innen
About
Self-supervised learning of graph neural networks (GNN) is in great need because of the widespread label scarcity issue in real-world graph/network data. Graph contrastive learning (GCL), by training GNNs to maximize the correspondence between the representations of the same graph in its different augmented forms, may yield robust and transferable GNNs even without using labels. However, GNNs trained by traditional GCL often risk capturing redundant graph features and thus may be brittle and provide sub-par performance in downstream tasks. Here, we propose a novel principle, termed adversarial-GCL (AD-GCL), which enables GNNs to avoid capturing redundant information during the training by optimizing adversarial graph augmentation strategies used in GCL. We pair AD-GCL with theoretical explanations and design a practical instantiation based on trainable edge-dropping graph augmentation. We experimentally validate AD-GCL by comparing with the state-of-the-art GCL methods and achieve performance gains of up-to 14% in unsupervised, 6% in transfer and 3% in semi-supervised learning settings overall with 18 different benchmark datasets for the tasks of molecule property regression and classification, and social network classification.Self-supervised learning of graph neural networks (GNN) is in great need because of the widespread label scarcity issue in real-world graph/network data. Graph contrastive learning (GCL), by training GNNs to maximize the correspondence between the representations of the same graph in its different augmented forms, may yield robust and transferable GNNs even without using labels. However, GNNs trained by traditional GCL often risk capturing redundant graph features and thus may be brittle and pr…
Organizer
NeurIPS 2021
Konto · 1,9k Follower:innen
About NeurIPS 2021
Neural Information Processing Systems (NeurIPS) is a multi-track machine learning and computational neuroscience conference that includes invited talks, demonstrations, symposia and oral and poster presentations of refereed papers. Following the conference, there are workshops which provide a less formal setting.
Like the format? Trust SlidesLive to capture your next event!
Professional recording and live streaming, delivered globally.
Sharing
Recommended Videos
Presentations on similar topic, category or speaker
Gradient Starvation: A Learning Proclivity in Neural Networks
Später ansehen
Ewigspeicher-Fortschrittswert: 1 = 0.1%
Structural Assumptions for Better Generalization in Reinforcement Learning
Später ansehen
Ewigspeicher-Fortschrittswert: 0 = 0.0%
Hybrid Regret Bounds for Combinatorial Semi-Bandits and Adversarial Linear Bandits
Später ansehen
Ewigspeicher-Fortschrittswert: 0 = 0.0%
Collective Intelligence of Army Ants and the Robots They Inspire
Später ansehen
Ewigspeicher-Fortschrittswert: 0 = 0.0%
Gradient-Driven Rewards to Guarantee Fairness in Collaborative Machine Learning
Später ansehen
Xinyi Xu, …
Ewigspeicher-Fortschrittswert: 0 = 0.0%
Heavy Ball Neural Ordinary Differential Equations
Später ansehen
Hedi Xia, …
Ewigspeicher-Fortschrittswert: 0 = 0.0%