Attacking and Defending Machine Learning Applications of Public Cloud
This addresses security issues for developers and users of cloud-based ML services, but appears incremental as it adapts existing SDL concepts to ML.
The paper tackles the problem of adversarial attacks on machine learning applications in public cloud services by proposing a Security Development Lifecycle (SDL) for ML, which helps developers reduce vulnerabilities and development costs in ML-as-a-service.
Adversarial attack breaks the boundaries of traditional security defense. For adversarial attack and the characteristics of cloud services, we propose Security Development Lifecycle for Machine Learning applications, e.g., SDL for ML. The SDL for ML helps developers build more secure software by reducing the number and severity of vulnerabilities in ML-as-a-service, while reducing development cost.