1: \begin{abstract}

2:    We propose a variance reduction framework for variational inference using the multi-level Monte Carlo (MLMC) method.

3: The proposed framework ``recycles'' parameters obtained from past update history in optimization and can be compatible with reparameterized gradient estimators. Our framework provides a novel optimization algorithm based on the stochastic gradient method and adaptively estimates the sample size for stochastic gradient estimation per level according to the ratio of the variance and computation cost in each iteration.

4: We also analyze the convergence of the gradient norm in the stochastic gradient method, the scale of the gradient estimator's variance, and the estimator's quality in each optimization step on the basis of the \textit{signal-to-noise} ratio. Finally, we experimentally evaluate the proposed method by comparing it with baseline methods on several benchmark data sets.

5: The results confirm that the proposed method achieves faster convergence and reduces the variance of the gradient estimator compared with the other methods.

6: \end{abstract}

7: