Research map / Graph learning / Graph self-supervised learning
Regularized Diffusion Process for Visual Retrieval
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| set Y to original affinity W instead of identity I (ablation) | RDP · ORL, K=11 | retrieval accuracy (%): 74.17 -> 75.08 | Helped |
| set Y to W instead of I (ablation) | RDP · MPEG-7, K=40 | retrieval accuracy (%): 93.77 -> 93.78 | No clear effect |
| set Y to W instead of I (ablation) | RDP · Ukbench | N-S score: 3.93 -> 3.94 | No clear effect |
| tensor product graph diffusion (RDP) instead of manifold ranking | manifold ranking · YALE B, K=15 | retrieval accuracy (%): 70.85 -> 78.24 | Helped |
| RDP instead of modified Manifold Ranking* (kNN graph, zero diagonal, Y=W) | manifold ranking · MPEG-7, K=40 | retrieval accuracy (%): 92.61 -> 93.78 | Helped |
| modified manifold ranking (kNN graph, zero diagonal, Y=W) vs standard | manifold ranking · MPEG-7, K=40 | retrieval accuracy (%): 89.26 -> 92.61 | Helped |
| different initializations of A(1) (ablation) | RDP · YALE B, K=15 | final retrieval accuracy | No clear effect |
| RDP vs TPG on Ukbench | tensor product graph diffusion · Ukbench | N-S score: 3.61 -> 3.94 | Helped |
Closest papers by meaning
- Hyperbolic Geometric Latent Diffusion Model for Graph Generation (ICML 2024 · similarity 0.86)
- Diffusion Models and the Manifold Hypothesis: Log-Domain Smoothing is Geometry Adaptive (NeurIPS 2025 · similarity 0.85)
- Locality Preserving Markovian Transition for Instance Retrieval (ICML 2025 · similarity 0.85)
- Text-to-Image Diffusion Models are Great Sketch-Photo Matchmakers (CVPR 2024 · similarity 0.85)
- READ: Retrieval-Enhanced Asymmetric Diffusion for Motion Planning (CVPR 2024 · similarity 0.85)
- Toward a Unified Geometry Understanding : Riemannian Diffusion Framework for Graph Generation and Prediction (NeurIPS 2025 · similarity 0.85)
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.