AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Architecture Matters: Uncovering Implicit Mechanisms in Graph Contrastive Learning

Xiaojun Guo, Yifei Wang, Zeming Wei, Yisen Wang

NeurIPS 2023 · Graph learning › Graph self-supervised learning · approach: Graph contrastive learning analysis · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
remove positive pairs (uniformity loss only) (ablation)SimCLR · CIFAR-10 linear probingaccuracy (%): 83.51 -> 27.51Hurt
remove positive pairs (uniformity loss only) (ablation)GRACE · Node classification avg, 5 datasetsaccuracy (%): 84.87 -> 83.43No clear effect
remove negative pairs (alignment loss only) (ablation)GRACE · Node classification avg, 5 datasetsaccuracy (%): 84.87 -> 30.37Hurt
remove negative pairs (alignment loss only) (ablation)GraphCL · Graph classification avg, 6 datasetsaccuracy (%): 71.29 -> 70.77No clear effect
remove negative pairs (alignment loss only) (ablation)AutoGCL · Graph classification avg, 6 datasetsaccuracy (%): 69.87 -> 70.60No clear effect
replace GCN encoder with MLP under no-positive training (ablation)GRACE · Node classification avg, 5 datasetsaccuracy (%): 77.07 -> 62.77Hurt
add ContraNorm to GCN encoder with alignment loss only (ablation)GRACE · Node classification avg, 5 datasetsaccuracy (%): 30.37 -> 82.96Improved
replace FM+EP augmentation with random Gaussian noise (ablation)GRACE · Node classification avg, 5 datasetsaccuracy (%): 84.87 -> 82.95No clear effect
remove all augmentations (ablation)GRACE · Node classification avg, 5 datasetsaccuracy (%): 84.87 -> 81.98No clear effect
replace default augmentation with Gaussian noise (ablation)SimCLR · CIFAR-10 linear probingaccuracy (%): 83.51 -> 36.56Hurt
remove positive pairs on heterophilic graphs (ablation)GRACE · Chameleon and Squirrel avgaccuracy (%): 40.88 -> 36.73Hurt
remove positive pairs (uniformity loss only) (ablation)GRACE · OGB-arxivaccuracy (%): 65.97 -> 65.49No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy