AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Aitchison Embeddings for Learning Compositional Graph Representations

Nikolaos Nakis, Chrysoula Kosma, Panagiotis Promponas, Michail Chatzianastasis, Giannis Nikolentzos

ICML 2026 · Graph learning › Graph self-supervised learning · approach: Semi-supervised node representation learning · 0 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 Aitchison/ILR geometry instead of Euclidean geometry on the simplex (ablation)AICoG latent distance model · AstroPh link prediction (D=64)AUC-ROC: .863 -> .976Improved
use Aitchison/ILR geometry instead of Euclidean geometry on the simplex (ablation)AICoG latent distance model · Cora node classification (D=64)Micro-F1: .442 -> .831Improved
learn the ILR basis jointly instead of fixed Helmert basis (ablation)AICoG · link prediction and node classificationAUC-ROC / Micro-F1: Cora D=64 .851 -> .852No clear effect
subcompositional restriction of trained model without retraining (ablation)AICoG (HB) · Cora link prediction (D=8)AUC-ROC: .837 -> .807No clear effect
subcompositional restriction of trained model without retraining (ablation)AICoG (HB) · Cora node classification (D=8)Micro-F1: .769 -> .672Hurt
AICoG instead of MMSBM for membership recoverysynthetic graphs · Bilinear-continuousl1 distance: 1.456 -> 0.922Improved
AICoG instead of MMSBM for interior membershipsCora · Cora membershipsentropy: 0.191 -> 1.064Improved
AICoG vs Euclidean Node2vec/Role2vec baselines on node classificationAICoG · Cora node classification (D=8)Micro-F1: .769 vs ROLE2VEC .783No clear effect
use AICoG instead of SLIM-RAAAICoG · DBLP link prediction (D=64)AUC-ROC: .962 vs .962No clear effect
AICoG vs SLIM-RAA at low dimensionAICoG · Cora node classification (D=8)Micro-F1: .628 -> .769Improved

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