Adaptive Optimizers Can Get a Policy Ranking Wrong: Momentum, Dependent Data, and a Two-Run Remedy
stat.ML, cs.LG, math.OC, math.ST, stat.TH
Submitted: 2026-10-31
Updated: 2026-10-31
Terminology
Sources
- When Does Dynamic Preconditioning Preserve the Polyak-Ruppert CLT? A Stabilization Threshold
- A Polyak-Ruppert Central Limit Theorem for SA-Adam with Momentum and Non-Convergent Adaptive Preconditioning
- Stochastic Approximation in a Markovian Framework Revisited: Lipschitz Continuity of the Poisson Equation
- Dopamine: A Research Framework for Deep Reinforcement Learning
- Convergence rates for the Adam optimizer
- Central limit theorem for the averaged Adam optimizer
- Adam symmetry theorem: characterization of the convergence of the stochastic Adam optimizer
- ODE approximation for the Adam algorithm: General and overparametrized setting
- Central Limit Theorems for Stochastic Approximation with controlled Markov chain dynamics
- Where Did My Optimum Go?: An Empirical Analysis of Gradient Descent Optimization in Policy Gradient Methods
- Bias and Extrapolation in Markovian Linear Stochastic Approximation with Constant Stepsizes
- The case for and against fixed step-size: Stochastic approximation algorithms in optimization and machine learning
- Regret Bounds and Experimental Design for Estimate-then-Optimize
- Self-Normalized Inference for Constant-Stepsize Temporal-Difference Learning under Markovian Sampling
Related papers
- Uniform-in-Time Weak Propagation-of-Chaos in Shallow Neural Networks
- Diffusion Models Observe Only Gradients: A Geometric Perspective on Score Matching Errors
- Quantifying Error Propagation and Model Collapse in Diffusion Models
- Behavior of prediction performance metrics with rare events
- Statistical Inference and Learning for Shapley Additive Explanations (SHAP)
- Control Variate Design for Score-Function Gradient Estimators in Importance-Weighted Variational Inference