The Impact of Uniform Inputs on Activation Sparsity and Energy-Latency Attacks in Computer Vision

arXiv:2403.18587v1 Announce Type: cross
Abstract: Resource efficiency plays an important role for machine learning nowadays. The energy and decision latency are two critical aspects to ensure a sustainable and practical application. Unfortunately, the energy consumption and decision latency are not robust against adversaries. Researchers have recently demonstrated that attackers can compute and submit so-called sponge examples at inference time to increase the energy consumption and decision latency of neural networks. In computer vision, the proposed strategy crafts inputs with less …

abstract application arxiv attacks computer computer vision consumption cs.cr cs.cv cs.lg decision efficiency energy impact inputs latency machine machine learning practical researchers resource efficiency robust role sparsity sustainable type uniform vision