Backdoor poisoning attacks against few-shot classifiers based on meta-learning
Open Access
- 1 January 2023
- journal article
- research article
- Published by Institute of Electronics, Information and Communications Engineers (IEICE) in Nonlinear Theory and Its Applications, IEICE
- Vol. 14 (2) , 491-499
Abstract
Few-shot classification is a classification made on the basis of very few samples, and meta-learning methods (also called “learning to learn”) are often employed to accomplish it. Research on poisoning attacks against meta-learning-based few-shot classifier has recently started to be investigated. While poisoning attacks aimed at disrupting the availability of the classifier during meta-testing have been studied in Xu et al. [1] and Oldewage et al. [2], backdoor poisoning in meta-testing has only been briefly explored by Oldewage et al. [2] under limited conditions. We formulate a backdoor poisoning attack on meta-learning-based few-shot classification in this study. We show that the proposed backdoor poisoning attack is effective against the few-shot classification using model-agnostic meta-learning (MAML) [3] through experiments.Keywords
This publication has 10 references indexed in Scilit:
- Meta-Learning in Neural Networks: A SurveyIEEE Transactions on Pattern Analysis and Machine Intelligence, 2021
- Yet Meta Learning Can Adapt Fast, it Can Also Break Easily2021
- STRIP2019
- Learning to Segment Skin Lesions from Noisy Annotations2019
- Training Medical Image Analysis Systems like Radiologists2018
- Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks2018
- Low Data Drug Discovery with One-Shot LearningACS Central Science, 2017
- Deep Residual Learning for Image Recognition2016
- Mastering the game of Go with deep neural networks and tree searchNature, 2016
- Human-level concept learning through probabilistic program inductionScience, 2015