通过分布对齐与信息流引导，生成更清晰的神经元可视化图像。
VITAL: More Understandable Feature Visualization through Distribution Alignment and Relevant Information Flow
- 用真实图像特征统计和网络信息流引导生成
- 可视化结果在多种模型上均优于现有方法
- 适合需要理解模型决策过程的研究者
神经网络广泛应用于复杂任务，尤其在高风险决策中，理解其推理过程至关重要，但现代深度网络难以解释。特征可视化（FV）是解析神经元响应信息的有效工具，通过生成人类可理解的图像来揭示神经元检测的内容。然而，现有方法常产生难以辨认的图像，包含重复模式和视觉伪影。为此，我们提出结合真实图像特征分布与相关网络信息流的引导机制，生成更具代表性的原型图像。该方法在多个网络架构中，无论是定性还是定量评价，均显著优于当前最优的特征可视化技术。它能有效揭示网络所依赖的信息，补充机制电路中关于信息编码位置的分析。代码已公开于：https://github.com/adagorgun/VITAL
原文摘要 · Abstract (English)
Neural networks are widely adopted to solve complex and challenging tasks. Especially in high-stakes decision-making, understanding their reasoning process is crucial, yet proves challenging for modern deep networks. Feature visualization (FV) is a powerful tool to decode what information neurons are responding to and hence to better understand the reasoning behind such networks. In particular, in FV we generate human-understandable images that reflect the information detected by neurons of interest. However, current methods often yield unrecognizable visualizations, exhibiting repetitive patterns and visual artifacts that are hard to understand for a human. To address these problems, we propose to guide FV through statistics of real image features combined with measures of relevant network flow to generate prototypical images. Our approach yields human-understandable visualizations that both qualitatively and quantitatively improve over state-of-the-art FVs across various architectures. As such, it can be used to decode which information the network uses, complementing mechanistic circuits that identify where it is encoded. Code is available at: https://github.com/adagorgun/VITAL
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。