Prototype Reinforcement for Few-Shot Learning

Xu, Lubin; Xie, Qian; Jiang, Baoqing; Zhang, Jiashuo

doi:10.1109/cac51589.2020.9326820

Cited by 1 publication

(1 citation statement)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tens of thousands of simulated tasks are randomly sampled from a distribution to train the model. The meta-learning based methods adopt a meta-training paradigm, which aims to make the model can fast adapt to novel class (Vinyals et al 2016;Gordon et al 2018;Lee et al 2019;Bertinetto et al 2018;Rusu et al 2019;Raghu et al 2020;Xu et al 2020). The metric-learning based methods directly compare the similarities of latent representations between query set and support set without fine-tuning on the support set, and use their relationship outputs to classify (Snell, Swersky, and Zemel 2017;Koch et al 2015;Sung et al 2018;Li et al 2019a;Doersch, Gupta, and Zisserman 2020;Zhang et al 2020;Wertheimer, Tang, and Hariharan 2021;Kang et al 2021).…”

Section: Related Workmentioning

confidence: 99%

Does Few-Shot Learning Suffer from Backdoor Attacks?

Liu,

Jia,

et al. 2024

AAAI

View full text Add to dashboard Cite

The field of few-shot learning (FSL) has shown promising results in scenarios where training data is limited, but its vulnerability to backdoor attacks remains largely unexplored. We first explore this topic by first evaluating the performance of the existing backdoor attack methods on few-shot learning scenarios. Unlike in standard supervised learning, existing backdoor attack methods failed to perform an effective attack in FSL due to two main issues. Firstly, the model tends to overfit to either benign features or trigger features, causing a tough trade-off between attack success rate and benign accuracy. Secondly, due to the small number of training samples, the dirty label or visible trigger in the support set can be easily detected by victims, which reduces the stealthiness of attacks. It seemed that FSL could survive from backdoor attacks. However, in this paper, we propose the Few-shot Learning Backdoor Attack (FLBA) to show that FSL can still be vulnerable to backdoor attacks. Specifically, we first generate a trigger to maximize the gap between poisoned and benign features. It enables the model to learn both benign and trigger features, which solves the problem of overfitting. To make it more stealthy, we hide the trigger by optimizing two types of imperceptible perturbation, namely attractive and repulsive perturbation, instead of attaching the trigger directly. Once we obtain the perturbations, we can poison all samples in the benign support set into a hidden poisoned support set and fine-tune the model on it. Our method demonstrates a high Attack Success Rate (ASR) in FSL tasks with different few-shot learning paradigms while preserving clean accuracy and maintaining stealthiness. This study reveals that few-shot learning still suffers from backdoor attacks, and its security should be given attention.

show abstract