AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Modular Flows: Differential Molecular Generation

Yogesh Verma, Samuel Kaski, Markus Heinonen, Vikas K Garg

NeurIPS 2022 · Graph learning › Graph self-supervised learning · approach: Molecular graph pretraining

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
replace E(3)-equivariant EGNN with non-equivariant 3-layer GCN (ablation)ModFlow (3D) · ZINC250Kvalidity %: 95.4 -> 90.3Hurt
replace E(3)-equivariant EGNN with non-equivariant 3-layer GCN (ablation)ModFlow (3D) · QM9validity %: 98.3 -> 93.3Hurt
use 3D instead of 2D coordinates (ablation)ModFlow EGNN · QM9validity %: 96.2 -> 98.3Improved
use 3D instead of 2D coordinates (ablation)ModFlow EGNN · ZINC250Kvalidity %: 94.8 -> 95.4Improved
add junction-tree representation to 3D coordinatesModFlow (3D-EGNN) · QM9validity %: 98.3 -> 99.1Improved
add junction-tree representation to 3D coordinatesModFlow (3D-EGNN) · ZINC250Kvalidity %: 95.4 -> 98.1Improved
add junction-tree representation to 2D coordinatesModFlow (2D-EGNN) · ZINC250Kvalidity %: 94.8 -> 97.4Improved
add junction-tree representation to 2D coordinatesModFlow (2D-EGNN) · QM9validity %: 96.2 -> 97.9Improved
use 3D instead of 2D coordinates (ablation)ModFlow EGNN · QM9uniqueness %: 99.5 -> 99.1No clear effect
use ModFlow vs MoFlow baseline (one-shot, no validity correction)ModFlow (JT-3D-EGNN) · ZINC250Kvalidity %: 50.3 -> 98.1Improved
use ModFlow vs GraphDF baseline on generation speedModFlow (2D-EGNN) · ZINC250Kseconds/molecule: 3.12 -> 0.46Improved
add junction-tree representation (slower decoding)ModFlow (2D-EGNN) · QM9seconds/molecule: 0.16 -> 0.21Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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