Forensicability of Deep Neural Network Inference Pipelines

TL;DR

This paper introduces methods to analyze the execution environment of deep neural network pipelines by detecting numerical deviations in outputs, enabling forensic identification of hardware platforms and machine-specific signatures.

Contribution

It presents novel techniques for inferring hardware and environment properties of neural network inference pipelines through numerical deviation analysis and boundary sample creation.

Findings

Proof-of-concept experiments on local and cloud machines validate the approach.

Numerical deviations can be used to identify hardware platforms.

Boundary samples effectively amplify deviations for machine differentiation.

Abstract

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.