Accelerating Stochastic Probabilistic Inference
This work addresses a bottleneck in probabilistic inference for researchers and practitioners, but it is incremental as it adapts existing second-order methods to SVI.
The paper tackles the problem of poor convergence rates in Stochastic Variational Inference (SVI) by proposing a second-order based approach, which is shown to be effective and efficient through empirical evaluations on synthetic and real datasets.
Recently, Stochastic Variational Inference (SVI) has been increasingly attractive thanks to its ability to find good posterior approximations of probabilistic models. It optimizes the variational objective with stochastic optimization, following noisy estimates of the natural gradient. However, almost all the state-of-the-art SVI algorithms are based on first-order optimization algorithm and often suffer from poor convergence rate. In this paper, we bridge the gap between second-order methods and stochastic variational inference by proposing a second-order based stochastic variational inference approach. In particular, firstly we derive the Hessian matrix of the variational objective. Then we devise two numerical schemes to implement second-order SVI efficiently. Thorough empirical evaluations are investigated on both synthetic and real dataset to backup both the effectiveness and efficiency of the proposed approach.