Tuning Stochastic Gradient Algorithms for Statistical Inference via Large-Sample Asymptotics
The tuning of stochastic gradient algorithms (SGAs) for optimization and sampling is often based on heuristics and trial-and-error rather than generalizable theory. We address this theory--practice gap by characterizing the large-sample statistical asymptotics of SGAs via a joint step-size--sample-size scaling limit. We show that iterate averaging with a large fixed step size is robust to the choice of tuning parameters and asymptotically has covariance proportional to that of the MLE sampling distribution. We also prove a Bernstein--von Mises-like theorem to guide tuning, including for generalized posteriors that are robust to model misspecification. Numerical experiments validate our results and recommendations in realistic finite-sample regimes. Our work lays the foundation for a systematic analysis of other stochastic gradient Markov chain Monte Carlo algorithms for a wide range of models.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Accurate Large-sample Uncertainty Quantification using Stochastic Gradient Markov Chain Monte Carlo
Tuning algorithms such as stochastic gradient descent (SGD) and stochastic gradient Langevin dynamics (SGLD) for approximate sampling and uncertainty quantification remains challenging, particularly in the practically re…
Statistical Inference for Polyak-Ruppert Averaged Zeroth-order Stochastic Gradient Algorithm
Statistical machine learning models trained with stochastic gradient algorithms are increasingly being deployed in critical scientific applications. However, computing the stochastic gradient in several such applications…
BIG-bench Machine Learningparameter estimationvalidAsymptotic and finite-sample properties of estimators based on stochastic gradients
Stochastic gradient descent procedures have gained popularity for parameter estimation from large data sets. However, their statistical properties are not well understood, in theory. And in practice, avoiding numerical i…
parameter estimationOnline Bootstrap Inference For Policy Evaluation in Reinforcement Learning
The recent emergence of reinforcement learning has created a demand for robust statistical inference methods for the parameter estimates computed using these algorithms. Existing methods for statistical inference in onli…
reinforcement-learningReinforcement LearningReinforcement Learning (RL)Decentralized Stochastic Gradient Langevin Dynamics and Hamiltonian Monte Carlo
Stochastic gradient Langevin dynamics (SGLD) and stochastic gradient Hamiltonian Monte Carlo (SGHMC) are two popular Markov Chain Monte Carlo (MCMC) algorithms for Bayesian inference that can scale to large datasets, all…
Bayesian Inferenceregression