提出可复用超参数的扩散优化器，加速图像重建且无需调参。
Principled Design of Diffusion-based Optimizers for Inverse Problems
- 通过重参数化实现跨任务超参数不变性，避免重复调优。
- 在图像重建、去模糊、超分辨率上提升速度与画质，最快提速4倍。
- 适合需要快速部署扩散模型的科研与工业应用。
基于分数的扩散模型在逆问题中表现卓越，但实际应用受限于推理时间长和超参数调优复杂。尽管预训练模型可跨任务复用，但推理时的噪声调度和后验采样权重通常需针对每种任务手动调整。本文提出原则性的重参数化方法，引入不变性，使同一组超参数可在多个问题间通用而无需重调。同时，在RED-diff框架基础上，构建OptDiff流水线，将后验采样转化为优化问题，简化调参流程，并融合凸优化工具加速推理。在图像重建、去模糊和超分辨率任务上的实验表明，该方法实现显著加速（最高4倍）并提升图像质量。
原文摘要 · Abstract (English)
Score-based diffusion models achieve state-of-the-art performance for inverse problems, but their practical deployment is hindered by long inference times and cumbersome hyperparameter tuning. While pretrained diffusion models can be reused across tasks without retraining, inference-time hyperparameters such as the noise schedule and posterior sampling weights typically require ad-hoc adjustment for each problem setup. We propose principled reparameterizations that induce invariances, allowing the same hyperparameters to be reused across multiple problems without re-tuning. In addition, building on the RED-diff framework, which reformulates posterior sampling as an optimization problem, we further develop the OptDiff pipeline. OptDiff provides a simplified tuning framework that facilitates the integration of convex optimization tools to accelerate inference. Experiments on image reconstruction, deblurring, and super-resolution show substantial speedups and improved image quality.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。