AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

GALA: Graph Diffusion-Based Alignment With Jigsaw for Source-Free Domain Adaptation

Junyu Luo, Yiyang Gu, Xiao Luo, Wei Ju, Zhiping Xiao, Yusheng Zhao, Jingyang Yuan, Ming Zhang

TPAMI 2024 · Graph learning › Graph self-supervised learning · approach: Graph domain adaptation · 38 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
add graph diffusion model alignment (ablation)GALA · Average over six datasets (ablation)accuracy (%): 54.7 -> 56.3Improved
add unbiased pseudo-labeling with curriculum learning (ablation)GALA · Average over six datasets (ablation)accuracy (%): 55.0 -> 56.3Improved
add consistency learning with graph jigsaw (ablation)GALA · Average over six datasets (ablation)accuracy (%): 55.5 -> 56.3Improved
add graph diffusion model alignment (ablation)GALA · ENZYMES (ablation)accuracy (%): 27.7 -> 30.7Improved
add unbiased pseudo-labeling with curriculum learning (ablation)GALA · ENZYMES (ablation)accuracy (%): 28.3 -> 30.7Improved
use GALA instead of PLUEPLUE · COX2 and BZR source-free (avg over 4 tasks)accuracy (%): 55.2 -> 60.5Improved
use GALA instead of PLUEPLUE · ENZYMES source-free (avg over 12 density-split tasks)accuracy (%): 26.1 -> 30.7Improved
use semi-supervised InfoGraph instead of source-only GCNGCN · PROTEINS source-free (avg over 12 density-split tasks)accuracy (%): 58.6 -> 59.4Mixed
raise initial pseudo-label threshold from 0.95 to 0.99 (ablation)GALA · average accuracy sensitivityaccuracy (%)Hurt
increase reconstruction timestamp t_recon beyond 0.1 (0.2 to 0.4) (ablation)GALA · average accuracy sensitivityaccuracy (%)Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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