Timezone: »
Poster
SGD with AdaGrad Stepsizes: Full Adaptivity with High Probability to Unknown Parameters, Unbounded Gradients and Affine Variance
Amit Attia · Tomer Koren
We study Stochastic Gradient Descent with AdaGrad stepsizes: a popular adaptive (self-tuning) method for first-order stochastic optimization. Despite being well studied, existing analyses of this method suffer from various shortcomings: they either assume some knowledge of the problem parameters, impose strong global Lipschitz conditions, or fail to give bounds that hold with high probability. We provide a comprehensive analysis of this basic method without any of these limitations, in both the convex and non-convex (smooth) cases, that additionally supports a general ``affine variance'' noise model and provides sharp rates of convergence in both the low-noise and high-noise regimes.
Author Information
Amit Attia (Tel Aviv University)
Tomer Koren (Tel Aviv University and Google)
More from the Same Authors
-
2023 Poster: Near-Optimal Algorithms for Private Online Optimization in the Realizable Regime »
Hilal Asi · Vitaly Feldman · Tomer Koren · Kunal Talwar -
2023 Poster: Improved Regret for Efficient Online Reinforcement Learning with Linear Function Approximation »
Uri Sherman · Tomer Koren · Yishay Mansour -
2023 Poster: Regret Minimization and Convergence to Equilibria in General-sum Markov Games »
Liad Erez · Tal Lancewicki · Uri Sherman · Tomer Koren · Yishay Mansour -
2021 Poster: Private Stochastic Convex Optimization: Optimal Rates in L1 Geometry »
Hilal Asi · Vitaly Feldman · Tomer Koren · Kunal Talwar -
2021 Oral: Private Stochastic Convex Optimization: Optimal Rates in L1 Geometry »
Hilal Asi · Vitaly Feldman · Tomer Koren · Kunal Talwar -
2021 Poster: Adversarial Dueling Bandits »
Aadirupa Saha · Tomer Koren · Yishay Mansour -
2021 Spotlight: Adversarial Dueling Bandits »
Aadirupa Saha · Tomer Koren · Yishay Mansour -
2021 Poster: Online Policy Gradient for Model Free Learning of Linear Quadratic Regulators with √T Regret »
Asaf Cassel · Tomer Koren -
2021 Poster: Stochastic Multi-Armed Bandits with Unrestricted Delay Distributions »
Tal Lancewicki · Shahar Segal · Tomer Koren · Yishay Mansour -
2021 Spotlight: Stochastic Multi-Armed Bandits with Unrestricted Delay Distributions »
Tal Lancewicki · Shahar Segal · Tomer Koren · Yishay Mansour -
2021 Spotlight: Online Policy Gradient for Model Free Learning of Linear Quadratic Regulators with √T Regret »
Asaf Cassel · Tomer Koren -
2021 Poster: Dueling Convex Optimization »
Aadirupa Saha · Tomer Koren · Yishay Mansour -
2021 Spotlight: Dueling Convex Optimization »
Aadirupa Saha · Tomer Koren · Yishay Mansour