Online Covariance Matrix Estimation in Stochastic Gradient Descent
This work addresses the need for statistical inference in online learning settings, providing a practical tool for researchers and practitioners using SGD, though it is incremental as it builds on existing asymptotic normality results.
The paper tackles the problem of quantifying variability in stochastic gradient descent (SGD) estimates by proposing a fully online estimator for the covariance matrix of averaged SGD iterates, establishing its consistency with a comparable convergence rate to offline methods and enabling asymptotically valid confidence intervals for model parameters.
The stochastic gradient descent (SGD) algorithm is widely used for parameter estimation, especially for huge data sets and online learning. While this recursive algorithm is popular for computation and memory efficiency, quantifying variability and randomness of the solutions has been rarely studied. This paper aims at conducting statistical inference of SGD-based estimates in an online setting. In particular, we propose a fully online estimator for the covariance matrix of averaged SGD iterates (ASGD) only using the iterates from SGD. We formally establish our online estimator's consistency and show that the convergence rate is comparable to offline counterparts. Based on the classic asymptotic normality results of ASGD, we construct asymptotically valid confidence intervals for model parameters. Upon receiving new observations, we can quickly update the covariance matrix estimate and the confidence intervals. This approach fits in an online setting and takes full advantage of SGD: efficiency in computation and memory.