Conformalized Multiple Testing after Data-dependent Selection

Authors: Xiaoning Wang, Yuyang Huo, Liuhua Peng, Changliang Zou

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Furthermore, we conduct numerical studies to showcase the effectiveness and validity of our procedures across various scenarios.
Researcher Affiliation Academia 1School of Statistics and Data Sciences, LPMC, KLMDASR and LEBPS, Nankai University, Tianjin, China 2School of Mathematics and Statistics, The University of Melbourne, Melbourne, Australia
Pseudocode Yes Algorithm 1 Selective conformal p-value with BH procedure (SCPV)
Open Source Code Yes We provide the data and code in the supplemental materials, including instructions in the zip file.
Open Datasets Yes Abalone[37]: contains easily obtainable measurements of abalone. The task is to predict the age of abalone from physical measurements.
Dataset Splits No The paper mentions a 'labeled/holdout data set Dc' and an 'unlabelled/test set Du', and their sizes, but does not explicitly specify a separate validation dataset split with proportions.
Hardware Specification Yes All the experiments were conducted on AMD Ryzen 7 5800H with Radeon Graphics processor with 16 Gb memory at a Lenovo personal computer and the R platform with version 4.2.1.
Software Dependencies Yes All the experiments were conducted on AMD Ryzen 7 5800H with Radeon Graphics processor with 16 Gb memory at a Lenovo personal computer and the R platform with version 4.2.1.
Experiment Setup Yes We fix the labeled data size n = 1, 200 and the unlabeled data size m = 1, 200. We fit the regression models ˆµ( ) on an additional labeled set with size 1, 200 using the random forest algorithm, implemented by R package random Forest with default parameters.