用扩散模型生成可调控的解释路径，让特征归因更准确清晰。
Diffusion Integrated Gradients: Controllable Path Generation for Flexible Feature Attribution
- 将路径生成转为条件生成问题，用扩散模型学习路径分布
- 通过引导采样实现用户可控的解释路径生成，提升解释质量
- 适合需要灵活、可控解释的AI应用开发者
基于路径的归因方法（如积分梯度）因其严格的公理性质和有效性被广泛采用，通过沿从基线到输入的路径积分梯度来归因模型预测。然而，归因路径的选择显著影响解释质量，现有方法依赖固定或手工设计的路径，常产生噪声大或失真的归因结果。为此，我们提出扩散积分梯度（DiffIG），将路径生成重新建模为条件生成问题。DiffIG首先训练扩散模型，学习由棒棒折分过程生成的路径分布，随后在采样过程中引入用户引导进行指导性采样。实验表明，DiffIG在定量指标上达到或超越现有路径方法，生成视觉上一致的解释。该工作为可灵活控制、推理时可调的可解释人工智能方法提供了新的生成视角。
原文摘要 · Abstract (English)
Path-based attribution methods such as Integrated Gradients (IG) are widely adopted for their strong axiomatic properties and effectiveness in attributing model predictions to input features by integrating gradients along a path from a baseline to the input. However, the choice of the attribution path largely affects the quality of explanations, and existing approaches rely on fixed or hand-crafted paths that often produce noisy or distorted attributions. To address this limitation, we propose Diffusion Integrated Gradients (DiffIG), a novel method that reformulates path generation as a conditional generative modeling problem. DiffIG first trains a diffusion model to learn a distribution over paths generated from a Stick-Breaking Process, then employs guided sampling to embed user guidance during the sampling procedure. We demonstrate that DiffIG quantitatively matches or outperforms existing path-based methods, achieving perceptually aligned explanations. This work introduces a new generative perspective for flexible, inference-time controllable Explainable Artificial Intelligence (XAI) methods.
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。