Unlocking Feature Visualization for Deeper Networks with MAgnitude Constrained Optimization
Unlocking Feature Visualization for Deeper Networks with MAgnitude Constrained Optimization
复制标题
DOI：
10.48550/arxiv.2306.06805
复制
发表时间：
2023-06
期刊：
影响因子：
--
通讯作者：
Thomas Fel;Thibaut Boissin;Victor Boutin;Agustin Picard;Paul Novello;Julien Colin;Drew Linsley;Tom Rousseau;Rémi Cadène;L. Gardes;Thomas Serre
中科院分区：
文献类型：
--
作者：
Thomas Fel;Thibaut Boissin;Victor Boutin;Agustin Picard;Paul Novello;Julien Colin;Drew Linsley;Tom Rousseau;Rémi Cadène;L. Gardes;Thomas Serre
特征可视化已经获得了相当大的普及，特别是在Olah等人有影响力的工作之后。2017年，将其确立为可解释性的重要工具。然而，由于依赖于生成可解释的图像的技巧，以及将其扩展到更深层次的神经网络方面的相应挑战，它的广泛采用一直受到限制。在这里，我们描述MACO，这是一种解决这些缺点的简单方法。其主要思想是通过优化相位谱来生成图像，同时保持幅度恒定，以确保生成的解释位于自然图像的空间中。我们的方法产生了显著更好的结果(无论是定性的还是定量的)，并为大型最先进的神经网络提供了高效和可解释的特征可视化。我们还表明，我们的方法展示了一种归因机制，允许我们用空间重要性来增强特征可视化。我们在一个新的比较要素可视化方法的基准上验证了我们的方法，并在https://serre-lab.github.io/Lens/.上发布了它对ImageNet数据集的所有类的可视化总体而言，我们的方法首次解锁了大型、最先进的深度神经网络的可视化功能，而不需要求助于任何参数先验图像模型。
Feature visualization has gained substantial popularity, particularly after the influential work by Olah et al. in 2017, which established it as a crucial tool for explainability. However, its widespread adoption has been limited due to a reliance on tricks to generate interpretable images, and corresponding challenges in scaling it to deeper neural networks. Here, we describe MACO, a simple approach to address these shortcomings. The main idea is to generate images by optimizing the phase spectrum while keeping the magnitude constant to ensure that generated explanations lie in the space of natural images. Our approach yields significantly better results (both qualitatively and quantitatively) and unlocks efficient and interpretable feature visualizations for large state-of-the-art neural networks. We also show that our approach exhibits an attribution mechanism allowing us to augment feature visualizations with spatial importance. We validate our method on a novel benchmark for comparing feature visualization methods, and release its visualizations for all classes of the ImageNet dataset on https://serre-lab.github.io/Lens/. Overall, our approach unlocks, for the first time, feature visualizations for large, state-of-the-art deep neural networks without resorting to any parametric prior image model.