Saved in:
Bibliographic Details
Main Authors: Fang, Yunxuan, Wang, Xinhe
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.09416
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909030164725760
author Fang, Yunxuan
Wang, Xinhe
author_facet Fang, Yunxuan
Wang, Xinhe
contents Hardware-aware training (HAT) is widely used to improve the robustness of neural networks on non-ideal AI accelerators, such as analog in-memory computing (IMC) systems. However, not all hardware-induced distortions are equally compensable by training. This paper presents a diagnostic framework that models hardware non-idealities as structured perturbations of the forward operator and evaluates their compatibility with gradient-based optimization. We analyze six representative perturbation classes--read noise, variability, drift, stuck-at faults, IR-drop, and ADC discretization--and identify three key diagnostics: gradient expectation consistency, bounded gradient variance, and non-degenerate sensitivity. Our results show a clear separation between perturbations that can be compensated by HAT and those that consistently break optimization. This provides practical guidance for hardware-software co-design, clarifying which non-idealities can be addressed at the training level and which require circuit-, architecture-, or calibration-level mitigation. This study should be interpreted as a controlled empirical analysis under vanilla forward-perturbation HAT, rather than as a universal theory of hardware-aware training.
format Preprint
id arxiv_https___arxiv_org_abs_2605_09416
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A Controlled Diagnostic Study of Hardware-Induced Distortions in Hardware-Aware Training
Fang, Yunxuan
Wang, Xinhe
Machine Learning
Hardware-aware training (HAT) is widely used to improve the robustness of neural networks on non-ideal AI accelerators, such as analog in-memory computing (IMC) systems. However, not all hardware-induced distortions are equally compensable by training. This paper presents a diagnostic framework that models hardware non-idealities as structured perturbations of the forward operator and evaluates their compatibility with gradient-based optimization. We analyze six representative perturbation classes--read noise, variability, drift, stuck-at faults, IR-drop, and ADC discretization--and identify three key diagnostics: gradient expectation consistency, bounded gradient variance, and non-degenerate sensitivity. Our results show a clear separation between perturbations that can be compensated by HAT and those that consistently break optimization. This provides practical guidance for hardware-software co-design, clarifying which non-idealities can be addressed at the training level and which require circuit-, architecture-, or calibration-level mitigation. This study should be interpreted as a controlled empirical analysis under vanilla forward-perturbation HAT, rather than as a universal theory of hardware-aware training.
title A Controlled Diagnostic Study of Hardware-Induced Distortions in Hardware-Aware Training
topic Machine Learning
url https://arxiv.org/abs/2605.09416