Sparse Data Augmentation for Optimization with Provable Guarantees

A theoretical machine-learning paper on data augmentation with provable optimization guarantees, with no direct AI-risk implications.

Auto-scored: Pure methodological research with no safety, misuse, or governance dimension.