AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Learning Invariant Graph Representations for Out-of-Distribution Generalization

Haoyang Li, Ziwei Zhang, Xin Wang, Wenwu Zhu

NeurIPS 2022 · Graph learning › Graph self-supervised learning · approach: Graph out-of-distribution generalization · 23 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
use GIL (latent-environment invariant learning) instead of ERMGIN backbone · SP-Motif r_test=1/3, r_train=0.8accuracy (%): 37.84 -> 51.24Improved
use GIL instead of DIRDIR · SP-Motif r_test=0.2, r_train=0.8accuracy (%): 39.92 -> 46.36Improved
use GIL instead of GSAT under strong spurious correlationGSAT · SP-Motif r_test=0.2, r_train=0.9accuracy (%): 33.87 -> 35.80Improved
use GIL instead of ERM on real graphsERM · MOLHIVROC-AUC (%): 76.20 -> 79.08Improved
use GIL instead of ERMERM · MNIST-75spaccuracy (%): 14.94 -> 21.94Improved
use GIL instead of DIRDIR · Graph-SST2accuracy (%): 83.29 -> 83.44No clear effect
wrap GIL around stronger backbone CINCIN · MOLHIVROC-AUC (%): 80.94 -> 81.15Improved
wrap GIL around stronger backbone HIGHIG · MOLHIVROC-AUC (%): 84.03 -> 84.23Improved
set number of inferred environments to ground truth (3) (ablation)GIL · SP-Motifaccuracy (%)Improved
use more than 2 inferred environments (ablation)GIL · MNIST-75spaccuracy (%)Hurt

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