Publication

Interpreting global perturbation robustness of image models using axiomatic spectral importance decomposition

Luo, Róisín (Jiaolin Luo)
McDermott, James
O'Riordan, Colm
Citation
Luo, Róisín (Jiaolin Luo), McDermott, James, & O’Riordan, Colm. (2024). Interpreting global perturbation robustness of image models using axiomatic spectral importance decomposition. Transactions on Machine Learning Research. https://openreview.net/pdf?id=uQYomAuo7M
Abstract
Perturbation robustness evaluates the vulnerabilities of models, arising from a variety of perturbations, such as data corruptions and adversarial attacks. Understanding the mechanisms of perturbation robustness is critical for global interpretability. We present a model-agnostic, global mechanistic interpretability method to interpret the perturbation robustness of image models. This research is motivated by two key aspects. First, previous global interpretability works, in tandem with robustness benchmarks, e.g., mean corruption error (mCE), are not designed to directly interpret the mechanisms of perturbation robustness within image models. Second, we notice that the spectral signal-to-noise ratios (SNR) of perturbed natural images exponentially decay over the frequency. This power-law-like decay implies that low-frequency signals are generally more robust than high-frequency signals—yet high classification accuracy cannot be achieved by low-frequency signals alone. By applying Shapley value theory, our method axiomatically quantifies the predictive powers of robust features and non-robust features within an information theory framework. Our method, dubbed as I-ASIDE (Image Axiomatic Spectral Importance Decomposition Explanation), provides a unique insight into model robustness mechanisms. We conduct extensive experiments over a variety of vision models pre-trained on ImageNet, including both convolutional neural networks (e.g., AlexNet, VGG, GoogLeNet/Inception-v1, Inception-v3, ResNet, SqueezeNet, RegNet, MnasNet, MobileNet, EfficientNet, etc.) and vision transformers (e.g., ViT, Swin Transformer, and MaxViT), to show that I-ASIDE can not only measure the perturbation robustness but also provide interpretations of its mechanisms.
Publisher
Transactions on Machine Learning Research
Publisher DOI
Rights
Attribution 4.0 International (CC BY 4.0)