arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.00921cs.LGcs.CRcs.MM
收藏

深度神经网络推理管道的可取证性

Forensicability of Deep Neural Network Inference Pipelines

University of Innsbruck(因斯布鲁克大学)

Alexander Schlögl, Tobias Kupek, Rainer Böhme

摘要:

我们提出了通过跟踪可观察输出中的特征数值偏差来推断机器学习管道的执行环境的属性的方法。在本地和云托管机器上获得的一系列概念验证实验的结果产生了可能的法医应用,例如识别用于产生深度神经网络预测的硬件平台。最后,我们引入边界样本,放大数值偏差,以便仅通过其预测标签来区分机器。

英文摘要:

We propose methods to infer properties of the execution environment of machine learning pipelines by tracing characteristic numerical deviations in observable outputs. Results from a series of proof-of-concept experiments obtained on local and cloud-hosted machines give rise to possible forensic applications, such as the identification of the hardware platform used to produce deep neural network predictions. Finally, we introduce boundary samples that amplify the numerical deviations in order to distinguish machines by their predicted label only.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑