Game-theoretic statistics and safe anytime-valid inference
Safe anytime-valid inference (SAVI) provides measures of statistical evidence and certainty—
e-processes for testing and confidence sequences for estimation—that remain valid at all …
e-processes for testing and confidence sequences for estimation—that remain valid at all …
Estimating means of bounded random variables by betting
I Waudby-Smith, A Ramdas - Journal of the Royal Statistical …, 2024 - academic.oup.com
We derive confidence intervals (CIs) and confidence sequences (CSs) for the classical
problem of estimating a bounded mean. Our approach generalizes and improves on the …
problem of estimating a bounded mean. Our approach generalizes and improves on the …
A unified recipe for deriving (time-uniform) PAC-Bayes bounds
We present a unified framework for deriving PAC-Bayesian generalization bounds. Unlike
most previous literature on this topic, our bounds are anytime-valid (ie, time-uniform) …
most previous literature on this topic, our bounds are anytime-valid (ie, time-uniform) …
Policy learning" without''overlap: Pessimism and generalized empirical Bernstein's inequality
This paper studies offline policy learning, which aims at utilizing observations collected a
priori (from either fixed or adaptively evolving behavior policies) to learn the optimal …
priori (from either fixed or adaptively evolving behavior policies) to learn the optimal …
Auditing fairness by betting
We provide practical, efficient, and nonparametric methods for auditing the fairness of
deployed classification and regression models. Whereas previous work relies on a fixed …
deployed classification and regression models. Whereas previous work relies on a fixed …
Jiayi Li, Yuantong Li and Xiaowu Dai's contribution to the Discussion of 'Estimating means of bounded random variables by betting'by Waudby-Smith and Ramdas
J Li, Y Li, X Dai - Journal of the Royal Statistical Society Series …, 2024 - academic.oup.com
1 Methods Consider a sequence of random variables (X t) t= 1∞, drawn from a distribution
P∈ P μ, where P μ represents the set of all distributions on [0, 1]∞. We assume EP [X t| X …
P∈ P μ, where P μ represents the set of all distributions on [0, 1]∞. We assume EP [X t| X …
Comparing sequential forecasters
Consider two forecasters, each making a single prediction for a sequence of events over
time. We ask a relatively basic question: how might we compare these forecasters, either …
time. We ask a relatively basic question: how might we compare these forecasters, either …
Art Owen's contribution to the Discussion of 'Estimating means of bounded random variables by betting'by Waudby-Smith and Ramdas
AB Owen - Journal of the Royal Statistical Society Series B …, 2024 - academic.oup.com
We derive confidence intervals (CIs) and confidence sequences (CSs) for the classical
problem of estimating a bounded mean. Our approach generalizes and improves on the …
problem of estimating a bounded mean. Our approach generalizes and improves on the …
Risk-limiting financial audits via weighted sampling without replacement
We introduce the notion of risk-limiting financial audits (RLFA): procedures that manually
evaluate a subset of $ N $ financial transactions to check the validity of a claimed assertion …
evaluate a subset of $ N $ financial transactions to check the validity of a claimed assertion …
Nonparametric extensions of randomized response for private confidence sets
I Waudby-Smith, S Wu… - … Conference on Machine …, 2023 - proceedings.mlr.press
This work derives methods for performing nonparametric, nonasymptotic statistical inference
for population means under the constraint of local differential privacy (LDP). Given bounded …
for population means under the constraint of local differential privacy (LDP). Given bounded …