1 code implementation • 19 Oct 2022 • Ian Waudby-Smith, Lili Wu, Aaditya Ramdas, Nikos Karampatziakis, Paul Mineiro
Importantly, our methods can be employed while the original experiment is still running (that is, not necessarily post-hoc), when the logging policy may be itself changing (due to learning), and even if the context distributions are a highly dependent time-series (such as if they are drifting over time).
no code implementations • 17 Feb 2022 • Ian Waudby-Smith, Zhiwei Steven Wu, Aaditya Ramdas
This work derives methods for performing nonparametric, nonasymptotic statistical inference for population means under the constraint of local differential privacy (LDP).
2 code implementations • 11 Mar 2021 • Ian Waudby-Smith, David Arbour, Ritwik Sinha, Edward H. Kennedy, Aaditya Ramdas
While the CLT approximates the distribution of a sample average by that of a Gaussian at a fixed sample size, we use strong invariance principles (stemming from the seminal 1970s work of Komlos, Major, and Tusnady) to uniformly approximate the entire sample average process by an implicit Gaussian process.
3 code implementations • 19 Oct 2020 • Ian Waudby-Smith, Aaditya Ramdas
This paper derives confidence intervals (CI) and time-uniform confidence sequences (CS) for the classical problem of estimating an unknown mean from bounded observations.
3 code implementations • NeurIPS 2020 • Ian Waudby-Smith, Aaditya Ramdas
We then present Hoeffding- and empirical-Bernstein-type time-uniform CSs and fixed-time confidence intervals for sampling WoR, which improve on previous bounds in the literature and explicitly quantify the benefit of WoR sampling.