提出属性解耦的引导方法，让生成的假设结果更真实可控。
Factored Classifier-Free Guidance
- 按因果图对各属性分别施加引导，避免全局控制导致的虚假变化。
- 在自然图像和医学图像上，显著提升反事实样本的逻辑一致性。
- 适用于各类扩散模型，特别适合需要精准干预的研究场景。
反事实生成旨在模拟因果干预下的真实假设结果。扩散模型凭借DDIM反演与条件生成结合分类器无关引导（CFG），成为该任务的重要工具。本文发现，现有CFG对所有属性采用统一引导强度，导致反事实推断中出现显著的虚假扰动。为此，提出属性解耦的分类器无关引导（FCFG），一种灵活且模型无关的引导机制，可依据因果图实现属性级控制。FCFG可无缝集成至CFG++、APG等先进引导方案。实验表明，FCFG在自然图像与医疗图像数据集上均显著提升反事实样本的公理一致性，有效缓解虚假放大效应，并增强反事实可逆性。
原文摘要 · Abstract (English)
Counterfactual generation aims to simulate realistic hypothetical outcomes under causal interventions. Diffusion models have emerged as a powerful tool for this task, combining DDIM inversion with conditional generation and classifier-free guidance (CFG). In this work, we identify a key limitation of CFG for counterfactual generation: it prescribes a global guidance scale for all attributes, leading to significant spurious changes in inferred counterfactuals. To mitigate this, we propose Factored Classifier-Free Guidance (FCFG), a flexible and model-agnostic guidance technique that enables attribute-wise control following a causal graph. FCFG complements recent advances in classifier-free guidance and can be seamlessly extended to advanced guidance schemes such as CFG++ and APG. Our experiments demonstrate that FCFG significantly improves the axiomatic soundness of inferred counterfactuals across both natural and medical image datasets, mitigating spurious amplification effects, and enhancing counterfactual reversibility.
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。