AI Paper Atlas

Research map / Graph learning / Knowledge and temporal graphs

NeuroRule: Bridging Vision and Logic with Differentiable Rule Induction

Muhammad Zarar, Mingzheng Zhang, Xiaowang Zhang, Zhiyong Feng

CVPR 2026 · Graph learning › Knowledge and temporal graphs · approach: Scene graph generation

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 symbolic single-step rule learning over neural-only MLP (ablation)NeuroRule · Visual Genome PredCls ablationR@50: 65.1 -> 67.4Improved
add multi-hop rule chaining (ablation)NeuroRule · Visual Genome PredCls ablationR@50: 67.4 -> 70.1Improved
add rule complexity regularization (ablation)NeuroRule · Visual Genome PredCls ablationR@50: 68.8 -> 70.1Improved
replace neural-only baseline with full NeuroRule (ablation)Mask2Former features · Visual Genome PredCls ablationR@100: 67.9 -> 72.9Improved
use NeuroRule over ViStruct on VG SGDetNeuroRule · Visual Genome SGDetAP50: 38.9 -> 42.3Improved
use NeuroRule over best baseline on PSGNeuroRule · PSG spatial relationsmean spatial accuracy: 78.4 -> 88.4Improved
use NeuroRule over ViStruct on Open-PSG zero-shotNeuroRule · Open-PSGOpen-Entity R@20: 35.2 -> 42.7Improved

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