Extragradient method improves AI generalization by finding flatter error minima

Original: Leveraging Extragradient for Effective Sharpness-Aware Minimization in Deep Learning

Source: arxiv.org ↗