期刊
INFORMATION SCIENCES
卷 632, 期 -, 页码 130-143出版社
ELSEVIER SCIENCE INC
DOI: 10.1016/j.ins.2023.02.086
关键词
Adversarial machine learning; Adversarial reprogramming; Neural networks; Transfer learning
Adversarial reprogramming enables repurposing of machine-learning models for different tasks. Recent research suggests that it can be used both for abusing models provided as a service and for improving transfer learning with limited training data. However, the factors that determine its success are still not well understood.
Adversarial reprogramming allows repurposing a machine-learning model to perform a different task. For example, a model trained to recognize animals can be reprogrammed to recognize digits by embedding an adversarial program in the digit images provided as input. Recent work has shown that adversarial reprogramming may not only be used to abuse machine-learning models provided as a service, but also beneficially, to improve transfer learning when training data is scarce. However, the factors affecting its success are still largely unexplained. In this work, we develop a first-order linear model of adversarial reprogramming to show that its success inherently depends on the size of the average input gradient, which grows when input gradients are more aligned, and when inputs have higher dimensionality. The results of our experimental analysis, involving fourteen distinct reprogramming tasks, show that the above factors are correlated with the success and the failure of adversarial reprogramming.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据