FiRe：用于视觉反事实解释的固定噪声细化
FiRe: Fixed-Noise Refinement for Visual Counterfactual Explanations
浏览论文内容
中文总结 AI 辅助
FiRe是一种用于视觉反事实解释的固定噪声细化框架，通过映射到固定噪声水平迭代细化、适配Pixel Mean Flow及三项专属控制机制，实现了更快推理、更低运算量与优质反事实质量。
中文摘要 AI 辅助
视觉反事实解释旨在通过现实且局部的编辑改变分类器的决策，同时保留与决策无关的内容。现有的基于DDPM的方法通常沿着较长的反向去噪轨迹执行分类器引导的编辑，不断变化的噪声水平使得语义可编辑性与空间控制难以平衡，且可编辑状态带有噪声，而目标分类器是在干净图像上训练的，因此这些方法要么需要代价高昂的递归去噪，要么需要低质量的单步估计来获得面向分类器的干净图像。我们提出FiRe，一种用于视觉反事实解释的固定噪声细化框架，FiRe不遵循反向去噪轨迹，而是将输入映射到固定噪声水平，并在该水平上迭代细化带噪声的状态；为了为分类器引导提供干净图像，FiRe将Pixel Mean Flow适配到视觉反事实解释任务，能够直接从带噪声状态预测干净图像；为了使固定噪声细化产生最小且局部的反事实编辑，FiRe引入了三个专属控制机制：动态双掩码策略、自适应引导和早停，分别确定编辑积累的位置、哪些变化会显现以及细化何时停止。在三个数据集的五项任务上进行的实验表明，与最近最强的基线相比，FiRe的在线推理速度提升了约3倍，浮点运算量（FLOPs）减少了8倍，同时获得了相当或达到当前最优水平的反事实质量。
英文摘要
Visual counterfactual explanations aim to change classifier decisions through realistic and localized edits while preserving decision-irrelevant content. Existing DDPM-based methods typically perform classifier-guided editing along a long reverse denoising trajectory. The changing noise levels make semantic editability and spatial control difficult to balance, and the editable state is noisy, whereas the target classifier is trained on clean images. As a result, these methods require either costly recursive denoising or low-quality one-step estimates to obtain classifier-facing clean images. We propose FiRe, a Fixed-noise Refinement framework for visual counterfactual explanations. Rather than following a reverse denoising trajectory, FiRe maps the input to a fixed noise level and iteratively refines the noisy state at that level. To provide clean images for classifier guidance, FiRe first adapts Pixel Mean Flow to visual counterfactual explanation, enabling direct clean-image prediction from noisy states. To make fixed-noise refinement produce minimal and localized counterfactual edits, FiRe introduces three FiRe-specific controls: a dynamic dual-mask strategy, adaptive guidance, and early stopping, which determine where edits accumulate, which changes become visible, and when refinement stops. Experiments on five tasks across three datasets show that, compared with the strongest recent baseline, FiRe achieves about 3$\times$ faster online inference and 8$\times$ fewer FLOPs while obtaining comparable or state-of-the-art counterfactual quality.
发表机构
- Technical University of Denmark（丹麦技术大学）
机构由 AI 辅助整理，请以论文原文为准。
