mlsanityData to 7 Oct 2026

Towards A Conceptually Simple Defensive Approach for Few-shot classifiers Against Adversarial Support Samples

arXiv 2110.12357 ↗First posted 24 Oct 2021cs.LG cs.CRcited by 0

Not on the map. Why?

Abstract (arXiv)

Few-shot classifiers have been shown to exhibit promising results in use cases where user-provided labels are scarce. These models are able to learn to predict novel classes simply by training on a non-overlapping set of classes. This can be largely attributed to the differences in their mechanisms as compared to conventional deep networks. However, this also offers new opportunities for novel attackers to induce integrity attacks against such models, which are not present in other machine learning setups. In this work, we aim to close this gap by studying a conceptually simple approach to defend few-shot classifiers against adversarial attacks. More specifically, we propose a simple attack-agnostic detection method, using the concept of self-similarity and filtering, to flag out adversarial support sets which destroy the understanding of a victim classifier for a certain class. Our extended evaluation on the miniImagenet (MI) and CUB datasets exhibit good attack detection performance, across three different few-shot classifiers and across different attack strengths, beating baselines. Our observed results allow our approach to establishing itself as a strong detection method for support set poisoning attacks. We also show that our approach constitutes a generalizable concept, as it can be paired with other filtering functions. Finally, we provide an analysis of our results when we vary two components found in our detection approach.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 25 of its 43 references that are in the map. First posted on arXiv 24 Oct 2021.

Builds on

25 of 43 references.

References · 25

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  4. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  5. 2013Intriguing properties of neural networkscited by 5,114
  6. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  7. 2017Prototypical Networks for Few-shot Learningcited by 3,001
  8. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,976
  9. 2016Matching Networks for One Shot Learningcited by 2,246
  10. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,510
  11. 2017Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networkscited by 1,325
  12. 2017Learning to Compare: Relation Network for Few-Shot Learningcited by 1,139
13 more
  1. 2016Practical Black-Box Attacks against Machine Learningcited by 893
  2. 2017Countering Adversarial Images using Input Transformationscited by 517
  3. 2017Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networkscited by 484
  4. 2018Meta-Transfer Learning for Few-Shot Learningcited by 306
  5. 2019Cross Attention Network for Few-shot Classificationcited by 218
  6. 2020Cross-Domain Few-Shot Classification via Learned Feature-Wise Transformationcited by 183
  7. 2016A Boundary Tilting Persepective on the Phenomenon of Adversarial Examplescited by 70
  8. 2019Dense Classification and Implanting for Few-Shot Learningcited by 56
  9. 2020TAFSSL: Task-Adaptive Feature Sub-Space Learning for few-shot classificationcited by 30
  10. 2020Learn2Perturb: an End-to-end Feature Perturbation Learning to Improve Adversarial Robustnesscited by 27
  11. 2018Adversarial Defense based on Structure-to-Signal Autoencoderscited by 10
  12. 2020Yet Meta Learning Can Adapt Fast, It Can Also Break Easilycited by 9
  13. 2020Detection of Adversarial Supports in Few-shot Classifiers Using Self-Similarity and Filteringcited by 1

Built on by

Not cited yet.