Efficient Data-Free Model Stealing with Label Diversity

arXiv:2404.00108v1 Announce Type: new
Abstract: Machine learning as a Service (MLaaS) allows users to query the machine learning model in an API manner, which provides an opportunity for users to enjoy the benefits brought by the high-performance model trained on valuable data. This interface boosts the proliferation of machine learning based applications, while on the other hand, it introduces the attack surface for model stealing attacks. Existing model stealing attacks have relaxed their attack assumptions to the data-free setting, while …

api applications arxiv benefits cs.cr data diversity free high interface machine machine learning opportunity performance proliferation query service stealing