2025
Stochastic Regret Guarantees for Online Zeroth- and First-Order Bilevel Optimization
NeurIPS 2025poster
Online bilevel optimization (OBO) is a powerful framework for machine learning problems where both outer and inner objectives evolve over time, requiring dynamic updates. Current OBO approaches rely on deterministic \textit{window-smoothed} regret minimization, which may not accurately reflect syste…