Weighted Uniform Sampling — a Monte Carlo Technique for Reducing Variance

By M. Powell, J. Swann. 1966

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Data-Efficient Off-Policy Policy Evaluation for Reinforcement Learning (2016, methodology): “The third notable trend is that when an accurate approximate model is available, WDR does not always outperform AM.”
Offline A/B Testing for Recommender Systems (2018, methodology): “…comprehensive study). such as Basic Importance Sampling (BIS, [8, 11]), Capped Importance Sampling (CIS, [2]), Normalised Importance Sampling (NIS, [21]) and Doubly Robust (DR, [6]) have been introduced to compute the expected reward of the tested technology π t based on logs collected on the…”
Importance Sampling for Fair Policy Selection (2017, methodology): “Although there is significant literature surrounding reducing the variance and mean squared error of off-policy policy evaluation methods that use IS-based estimators (Powell and Swann, 1966; Dudı́k et al., 2011; Jiang and Li, 2016; Thomas and Brunskill, 2017), other challenges associated with off-p”
Safe Reinforcement Learning (2015, methodology): “It is well known that IS and PDIS are unbiased estimators and that IS, PDIS, and WIS are consistent estimators if using historical data from a single behavior policy (Powell and Swann, 1966, Precup, 2000, Rubinstein and Kroese, 2007).”
A comparative study of counterfactual estimators (2017, background): “[10] MJD Powell and J Swann.”