Adversarially Robust Few-Shot Learning: A Meta-Learning Approach
arXiv:1910.00982
Abstract
Previous work on adversarially robust neural networks for image classification requires large training sets and computationally expensive training procedures. On the other hand, few-shot learning methods are highly vulnerable to adversarial examples. The goal of our work is to produce networks which both perform well at few-shot classification tasks and are simultaneously robust to adversarial examples. We develop an algorithm, called Adversarial Querying (AQ), for producing adversarially robust meta-learners, and we thoroughly investigate the causes for adversarial vulnerability. Moreover, our method achieves far superior robust performance on few-shot image classification tasks, such as Mini-ImageNet and CIFAR-FS, than robust transfer learning.
arXiv commentAccepted to NeurIPS 2020
References in corpus (13)
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
- Prototypical Networks for Few-shot Learning
- Towards Deep Learning Models Resistant to Adversarial Attacks
- Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
- Theoretically Principled Trade-off between Robustness and Accuracy
- Robust Physical-World Attacks on Deep Learning Models
- Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Models
- Learning to Remember Rare Events
- The Robust Manifold Defense: Adversarial Training using Generative Models
- Adversarial Meta-Learning
- Data Augmentation for Meta-Learning
- Adversarial attacks on Copyright Detection Systems
- Unraveling Meta-Learning: Understanding Feature Representations for Few-Shot Tasks
Cited by in corpus (7)
- Enhancing Few-Shot Image Classification through Learnable Multi-Scale Embedding and Attention Mechanisms
- Data Augmentation for Meta-Learning
- On Fast Adversarial Robustness Adaptation in Model-Agnostic Meta-Learning
- Unraveling Meta-Learning: Understanding Feature Representations for Few-Shot Tasks
- Signal Transformer: Complex-valued Attention and Meta-Learning for Signal Recognition
- Adversarial Monte Carlo Meta-Learning of Optimal Prediction Procedures
- On sensitivity of meta-learning to support data