Research map / Image generation / Fast diffusion sampling
Diffusion-based learning framework for Constrained Nonconvex Optimization with Weighted Bootstrapped Refinement
Shutong Ding, Yimiao Zhou, Ke Hu, Xi Yao, Junchi Yan, Xiaoying Tang, Ye Shi
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add reset operation (ablation) | DiOpt · QPSR | feasibility (%): 70.03 -> 81.19 | Improved |
| add reset operation (ablation) | DiOpt · CQP | feasibility (%): 83.79 -> 87.07 | Improved |
| add look-up table of best solutions (ablation) | DiOpt · QPSR | optimality gap (%): 112.8 -> 6.69 | Improved |
| add look-up table of best solutions (ablation) | DiOpt · QPSR | average constraint violation: 0.120 -> 0.008 | Improved |
| add bootstrapped refinement stage on top of supervised diffusion (RectFlow) | DiOpt · QPSR | feasibility (%): 0.00 -> 81.87 | Improved |
| add bootstrapped refinement stage on top of supervised diffusion (RectFlow) | DiOpt · ACOPF118 | feasibility (%): 0.00 -> 84.33 | Improved |
| DiOpt versus DC3 on ACOPF57 | DiOpt · ACOPF57 | feasibility (%): 94.00 -> 93.33 | Hurt |
| increase diffusion steps T to 100 (ablation) | DiOpt · QPSR, CQP, ACOPF, Retargeting | optimality gap | Improved |
| increase T from 5 to 100 (training time) (ablation) | DiOpt · QPSR | training time (hours): 3.09 -> 53.35 | Hurt |
| reduce supervised ratio r_s to 0.02 (ablation) | DiOpt · QP and QPSR (dim 100) | objective performance | Hurt |
| increase training samples K_t (ablation) | DiOpt · QPSR | objective vs feasibility | Mixed |
| set noise level eta=0 (no noise injection) (ablation) | DiOpt · toy example | sample diversity | Hurt |
Closest papers by meaning
- Fast T2T: Optimization Consistency Speeds Up Diffusion-Based Training-to-Testing Solving for Combinatorial Optimization (NeurIPS 2024 · similarity 0.87)
- A Gradient Guided Diffusion Framework for Chance Constrained Programming (NeurIPS 2025 · similarity 0.86)
- Moreau Envelope for Nonconvex Bi-Level Optimization: A Single-Loop and Hessian-Free Solution Strategy (ICML 2024 · similarity 0.86)
- Gradient Guidance for Diffusion Models: An Optimization Perspective (NeurIPS 2024 · similarity 0.86)
- Posterior Inference with Diffusion Models for High-dimensional Black-box Optimization (ICML 2025 · similarity 0.86)
- Variational Trajectory Optimization of Anisotropic Diffusion Schedules (NeurIPS 2026 · 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.
