Semantics Preserving Adversarial Learning
arXiv:1903.03905
Abstract
While progress has been made in crafting visually imperceptible adversarial examples, constructing semantically meaningful ones remains a challenge. In this paper, we propose a framework to generate semantics preserving adversarial examples. First, we present a manifold learning method to capture the semantics of the inputs. The motivating principle is to learn the low-dimensional geometric summaries of the inputs via statistical inference. Then, we perturb the elements of the learned manifold using the Gram-Schmidt process to induce the perturbed elements to remain in the manifold. To produce adversarial examples, we propose an efficient algorithm whereby we leverage the semantics of the inputs as a source of knowledge upon which we impose adversarial constraints. We apply our approach on toy data, images and text, and show its effectiveness in producing semantics preserving adversarial examples which evade existing defenses against adversarial attacks.
References in corpus (21)
- Auto-Encoding Variational Bayes
- Conditional Image Synthesis With Auxiliary Classifier GANs
- Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
- Provable defenses against adversarial examples via the convex outer adversarial polytope
- Synthesizing Robust Adversarial Examples
- Certified Defenses against Adversarial Examples
- Understanding Neural Networks through Representation Erasure
- Generating Natural Adversarial Examples
- A large annotated corpus for learning natural language inference
- Adversarial Examples Are Not Bugs, They Are Features
- Understanding Adversarial Training: Increasing Local Stability of Neural Nets through Robust Optimization
- InfoVAE: Information Maximizing Variational Autoencoders
- Adversarial Examples for Evaluating Reading Comprehension Systems
- Early Visual Concept Learning with Unsupervised Deep Learning
- Universal adversarial perturbations
- Explorations in Homeomorphic Variational Auto-Encoding
- Variational Gaussian Dropout is not Bayesian
- Stein Variational Adaptive Importance Sampling
- VAE Learning via Stein Variational Gradient Descent
- Symmetric Variational Autoencoder and Connections to Adversarial Learning
- On Certifying Non-uniform Bound against Adversarial Attacks