Joint control variate for faster black-box variational inference
This work addresses a specific bottleneck in variational inference for machine learning practitioners, offering an incremental improvement over existing methods.
The paper tackled the problem of high variance in gradient estimators for black-box variational inference, which arises from both data subsampling and Monte Carlo sampling, by proposing a joint control variate that reduces variance from both sources, leading to faster optimization in several applications.
Black-box variational inference performance is sometimes hindered by the use of gradient estimators with high variance. This variance comes from two sources of randomness: Data subsampling and Monte Carlo sampling. While existing control variates only address Monte Carlo noise, and incremental gradient methods typically only address data subsampling, we propose a new "joint" control variate that jointly reduces variance from both sources of noise. This significantly reduces gradient variance, leading to faster optimization in several applications.