aiqbench analysis
This research introduces an autonomic framework that optimizes adversarial robustness in deep learning by dynamically balancing cost, latency, and hardware selection. By leveraging accelerated failure time models within a MAPE-K control loop, the study demonstrates that cheaper GPUs like the Nvidia L4 can significantly outperform expensive alternatives such as the V100 for specific security-critical tasks. The findings highlight inference latency as a superior predictor of robustness compared to raw training duration or hardware specs, offering practical guidance for deploying secure AI systems on constrained budgets.
Survival of the Cheapest: Cost-Aware Hardware Adaptation for Adversarial Robustness
arXiv:2409.07609v3 Announce Type: replace-cross Abstract: Deploying adversarially robust machine learning systems requires continuous trade-offs between robustness, cost, and latency. We present an autonomic decision-support framework providing a quantitative foundation for adaptive hardware selection and hyper-parameter tuning in cloud-native deep learning. The framework applies accelerated failure time (AFT) models to quantify the effect of hardware choice, batch size, epochs, and validation accuracy on model survival time. This framework can be naturally integrated into an autonomic control loop (monitor--analyse--plan--execute, MAPE-K), where system metrics such as cost, robustness, and latency are continuously evaluated and used to adapt model configurations and hardware selection. Experiments across three GPU architectures confirm the framework is both sound and cost-effective: the Nvidia L4 yields a 20% increase in adversarial survival time while costing 75% less than the V100, demonstrating that expensive hardware does not necessarily improve robustness. The analysis further reveals that model inference latency is a stronger predictor of adversarial robustness than training time or hardware configuration.
Abstract:Deploying adversarially robust machine learning systems requires continuous trade-offs between robustness, cost, and latency. We present an autonomic decision-support framework providing a quantitative foundation for adaptive hardware selection and hyper-parameter tuning in cloud-native deep learning. The framework applies accelerated failure time (AFT) models to quantify the effect of hardware choice, batch size, epochs, and validation accuracy on model survival time. This framework can be naturally integrated into an autonomic control loop (monitor--analyse--plan--execute, MAPE-K), where system metrics such as cost, robustness, and latency are continuously evaluated and used to adapt model configurations and hardware selection. Experiments across three GPU architectures confirm the framework is both sound and cost-effective: the Nvidia L4 yields a 20% increase in adversarial survival time while costing 75% less than the V100, demonstrating that expensive hardware does not necessarily improve robustness. The analysis further reveals that model inference latency is a stronger predictor of adversarial robustness than training time or hardware configuration.
Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Applications (stat.AP)
arXiv:2409.07609 [cs.CR]
(or arXiv:2409.07609v3 [cs.CR] for this version)
Full text under the publisher's open licence · 178 words
