SUPREME: A Multi-GPU Framework for Reproducible Image Unlearning Method Evaluation
Researchers have developed SUPREME, an open-source framework designed to streamline the evaluation of machine unlearning methods for image data. This multi-GPU system addresses the computational expense of unlearning by distributing training, unlearning, and evaluation processes across multiple accelerators. SUPREME features a registry-based design for easy integration of new components and has been demonstrated on a face recognition task using standard models. AI
IMPACT Enables more efficient and reproducible evaluation of machine unlearning techniques, potentially accelerating progress in data privacy for AI models.