Bolstering Stochastic Gradient Descent with Model Building

arXiv:2111.07058v4 Announce Type: replace
Abstract: Stochastic gradient descent method and its variants constitute the core optimization algorithms that achieve good convergence rates for solving machine learning problems. These rates are obtained especially when these algorithms are fine-tuned for the application at hand. Although this tuning process can require large computational costs, recent work has shown that these costs can be reduced by line search methods that iteratively adjust the step length. We propose an alternative approach to stochastic line search …

abstract algorithms application arxiv building computational convergence core costs cs.lg good gradient machine machine learning optimization process stochastic type variants work