AINeutralarXiv – CS AI · 18h ago6/10
🧠
SVRG and Beyond via Posterior Correction
Researchers have established a fundamental connection between Stochastic Variance Reduced Gradient (SVRG), a decade-old optimization method, and Bayesian posterior correction techniques. This theoretical breakthrough enables the derivation of novel SVRG extensions using flexible exponential-family posteriors, including Newton-like and Adam-like variants that improve training efficiency.