AI Paper Atlas

Research map / Graph learning / GNN architectures

Deep Constraint-Based Propagation in Graph Neural Networks

Matteo Tiezzi, Giuseppe Marra, Stefano Melacci, Marco Maggini

TPAMI 2021 · Graph learning › GNN architectures · approach: Simple and scalable GNNs · 17 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
train with Lagrangian constraint-based propagation instead of iterative fixed-point GNN* trainingGNN* · Clique localizationaccuracy (%): 91.86 -> 93.07Improved
train with Lagrangian constraint-based propagation instead of iterative fixed-point GNN* trainingGNN* · Subgraph matchingaccuracy (%): 95.86 -> 96.30No clear effect
use squared G instead of lin G (ablation)LP-GNN · Clique localizationaccuracy (%): 84.61 -> 93.07Improved
use unilateral abs G instead of bilateral lin G (ablation)LP-GNN · Clique localizationaccuracy (%): 84.61 -> 88.80Improved
increase epsilon from 0.00 to 0.10 for abs G (ablation)LP-GNN · Clique localizationaccuracy (%): 88.80 -> 85.88Hurt
use multiple state layers instead of a single layer (ablation)LP-GNN · NCI1accuracy (%): 68.4 -> 74.9Improved
use multiple state layers instead of a single layer (ablation)LP-GNN · IMDB-Baccuracy (%): 65.3 -> 76.2Improved
use multiple state layers instead of a single layer (ablation)LP-GNN · PTCaccuracy (%): 64.4 -> 67.9Improved
LP training vs original GNN* trainingGNN* · NCI1accuracy (%): 51.5 -> 68.4Improved
LP training vs original GNN* trainingGNN* · IMDB-Baccuracy (%): 60.9 -> 65.3Improved
use LP-GNN with one state layer instead of GINGIN · IMDB-B (1 state layer)accuracy (%): 52 -> 65.3Improved
use 2 state layers instead of 5 (ablation)LP-GNN · IMDB-Baccuracy (%): 76.2 -> 73.7Mixed

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