让扩散模型生成更准：通过时间一致性提升预测精度
DiFA: Inference-Time Forward-Process Alignment for Diffusion Models
- 将生成过程视为状态估计，利用历史预测构建时间一致性共识
- 在CIFAR-10和ImageNet上显著提升FID、IS、FD-DINOv2指标
- 无需训练，适合追求高保真图像生成的研究者与应用
当前扩散模型的推理框架将生成视为数值积分问题，将模型视为精确估计器，忽略了去噪过程中的固有统计不确定性。本文提出无需训练的前向过程对齐扩散预测（DiFA），将推理时的数据预测优化重构为序列状态估计问题。不同于仅用于数值积分的历史输出复用，DiFA将逆向轨迹上的迭代预测视为相关观测，基于结构一致性和噪声水平兼容性构建前向对齐的时间共识。为抑制时间共识导致的过度平滑，引入偏差引导机制以自适应保留残差细节。实验表明，DiFA在CIFAR-10和ImageNet上均显著提升FID、IS及FD-DINOv2指标，证明对齐推理与前向统计结构可大幅提高生成保真度。
原文摘要 · Abstract (English)
The prevailing inference framework for diffusion models formulates generation fundamentally as a problem of numerical integration. This perspective casts the model as an exact estimator, neglecting the inherent statistical uncertainty of the denoising process. In this work, we propose Forward-Process Aligned Diffusion prediction (\textbf{DiFA}), a training-free framework that reframes inference-time data prediction refinement as a sequential state estimation problem. Rather than reusing past outputs solely for numerical integration, DiFA treats iterative data predictions along the reverse trajectory as correlated observations to build a forward-aligned temporal consensus. Inspired by Kalman filtering, this consensus aggregates historical predictions according to structural consistency and noise-level compatibility. To counteract the over-smoothing tendency of temporal consensus, we introduce a deviation guidance mechanism to adaptively preserve residual details. Empirically, DiFA yields significant improvements on CIFAR-10 and ImageNet across the evaluated metrics, including FID, IS, and FD-DINOv2, demonstrating that aligning inference with the forward statistical structure substantially improves generative fidelity.
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。