Non-Asymptotic Convergence of Stochastic Iterative Algorithms: A Lyapunov Framework
Researchers have published a paper detailing a Lyapunov-based framework for analyzing the finite-time convergence of stochastic iterative algorithms. This approach uses generalized Moreau envelopes as universal Lyapunov functions, applicable across various norms and noise types. The framework provides mean-square convergence guarantees and extends to algorithms like stochastic gradient descent and reinforcement learning methods such as Q-learning and temporal-difference learning. AI
IMPACT Provides a unified framework for analyzing convergence in reinforcement learning and other stochastic algorithms.