Adversarial Attacks on the Interpretation of Neuron Activation Maximization
The internal functional behavior of trained Deep Neural Networks is notoriously difficult to interpret. Activation-maximization approaches are one set of techniques used to interpret and analyze trained deep-learning models. These consist in finding inputs that maximally activate a given neuron or feature map. These inputs can be selected from a data set or obtained by optimization. However, interpretability methods may be subject to being deceived. In this work, we consider the concept of an adversary manipulating a model for the purpose of deceiving the interpretation. We propose an optimization framework for performing this manipulation and demonstrate a number of ways that popular activation-maximization interpretation techniques associated with CNNs can be manipulated to change the interpretations, shedding light on the reliability of these methods.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Representation Understanding via Activation Maximization
Understanding internal feature representations of deep neural networks (DNNs) is a fundamental step toward model interpretability. Inspired by neuroscience methods that probe biological neurons using visual stimuli, rece…
Bounded Adversarial Attack on Deep Content Features
We propose a novel adversarial attack targeting content features in some deep layer, that is, individual neurons in the layer. A naive method that enforces a fixed value/percentage bound for neuron activation values …
Adversarial AttackReal-time Detection of Practical Universal Adversarial Perturbations
Universal Adversarial Perturbations (UAPs) are a prominent class of adversarial examples that exploit the systemic vulnerabilities and enable physically realizable and robust attacks against Deep Neural Networks (DNNs). …
Blockingimage-classificationImage Classificationobject-detection+1TRM-UAP: Enhancing the Transferability of Data-Free Universal Adversarial Perturbation via Truncated Ratio Maximization
Aiming at crafting a single universal adversarial perturbation (UAP) to fool CNN models for various data samples, universal attack enables a more efficient and accurate evaluation for the robustness of CNN models. Ea…
Interpreting Adversarial Examples by Activation Promotion and Suppression
It is widely known that convolutional neural networks (CNNs) are vulnerable to adversarial examples: images with imperceptible perturbations crafted to fool classifiers. However, interpretability of these perturbations i…
Adversarial Robustness