SysNoise: Exploring and Benchmarking Training-Deployment System Inconsistency

TL;DR

This paper introduces SysNoise, a new type of noise caused by system mismatches during deep learning deployment, and provides a comprehensive benchmark showing its impact across various models and tasks, highlighting the need for robustness research.

Contribution

It is the first to identify and classify SysNoise, and to build a benchmark measuring its impact on over 20 models across multiple tasks.

Findings

SysNoise significantly affects model robustness.

Common mitigation techniques have limited effectiveness.

The work opens new research directions in deployment system robustness.

Abstract

Extensive studies have shown that deep learning models are vulnerable to adversarial and natural noises, yet little is known about model robustness on noises caused by different system implementations. In this paper, we for the first time introduce SysNoise, a frequently occurred but often overlooked noise in the deep learning training-deployment cycle. In particular, SysNoise happens when the source training system switches to a disparate target system in deployments, where various tiny system mismatch adds up to a non-negligible difference. We first identify and classify SysNoise into three categories based on the inference stage; we then build a holistic benchmark to quantitatively measure the impact of SysNoise on 20+ models, comprehending image classification, object detection, instance segmentation and natural language processing tasks. Our extensive experiments revealed that SysNoise could bring certain impacts on model robustness across different tasks and common mitigations like data augmentation and adversarial training show limited effects on it. Together, our findings open a new research topic and we hope this work will raise research attention to deep learning deployment systems accounting for model performance. We have open-sourced the benchmark and framework at https://modeltc.github.io/systemnoise_web.