GRIP2: A Robust and Powerful Deep Knockoff Method for Feature Selection

Bob Junyi Zou, Lu Tian · 2026

Read GRIP2: A Robust and Powerful Deep Knockoff Method for Feature Selection by Bob Junyi Zou, Lu Tian (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-00218.txt.

Description

Bob Junyi Zou and Lu Tian propose GRIP2, a deep knockoff method for feature selection with strict false discovery control. Picking truly predictive traits is hard when traits correlate strongly, signals are weak, and noise is high. GRIP2 builds a feature importance statistic by adding up first layer activity over a two dimensional regularization surface. Block stochastic sampling approximates that integral in one training run. The statistics are built to be antisymmetric, giving finite sample FDR control, and tests show they hold up under correlation and noise.

Catalog record

CREATORBob Junyi Zou, Lu Tian
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE69090
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.00218

Browse the full library · About the Tower of Babel