发表机构
Saarland University; Zuse School ELIZA（萨尔大学; 楚塞学校ELIZA）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出神经回声工具，将经典信号处理的脉冲响应等概念推广到神经网络，可解释各类网络行为，以DnCNN为例验证其能复现经典去噪器关键概念，搭建信号处理与可解释AI的桥梁。
AI 中文摘要
我们引入神经回声作为理解神经网络行为的工具，它将脉冲响应、扩散回声和滤波器回声这类基于模型的概念推广到基于学习的方法中。神经回声为神经网络提供局部、空间自适应的脉冲响应和滤波器核（即所谓的“回声”），这些回声依赖于输入图像，可通过仿射映射可视化，以理解网络学到的动态。神经回声搭建了经典信号处理与现代可解释AI之间的桥梁，通用性极强，可应用于图像到图像转换网络、分类网络，包含卷积或全连接结构、前馈或循环类型，还包括现代Transformer网络；且不要求网络可微，在可微情况下，神经回声包含基于网络雅可比矩阵的概念（如显著性图、对抗扰动分析）作为特例。作为解释该框架的简单示例，我们推导了去噪卷积神经网络（DnCNN）的神经回声，实验表明该网络基于像素的空间和灰度值距离对其加权，这不仅阐明了其行为，还显示它能复现双边滤波等经典基于模型的去噪器的关键概念。
英文摘要
We introduce the neural echo as a tool for understanding the behavior of neural networks. It generalizes the model-based concepts of impulse responses, diffusion echoes, and filter echoes to learning-based methods. It provides local, space-adaptive impulse responses and filter kernels for a neural network, its so-called echoes. These echoes depend on the input image and can be visualized to understand the learned dynamics of the network via an affine mapping. Neural echoes build a bridge from classical signal processing to modern explainable AI. They are very general and can be applied to both image-to-image and classification networks, with convolutional or fully connected structure, of feedforward or recurrent type, including modern transformer networks. Network differentiability is not required. In the differentiable case, neural echoes comprise concepts based on the network Jacobian, such as saliency maps and the analysis of adversarial perturbations, as special instances. As a simple blueprint to explain our framework, we derive neural echoes for the denoising convolutional neural network (DnCNN). Our experiments suggest that this network weights pixels based on their spatial and gray value distances. This not only clarifies its behavior, but also shows that it can reproduce key concepts of classical model-based denoisers such as bilateral filtering.
