Outcome-Aware Spectral Feature Learning for Instrumental Variable Regression

Dimitri Meunier [email protected], Jakub Wornbard [email protected], Vladimir R. Kostic [email protected] Antoine MoulinAlek Fröhlich Karim Lounici, Massimiliano Pontil Arthur Gretton · 2025

Read Outcome-Aware Spectral Feature Learning for Instrumental Variable Regression by Dimitri Meunier [email protected], Jakub Wornbard [email protected], Vladimir R. Kostic [email protected] Antoine MoulinAlek Fröhlich Karim Lounici, Massimiliano Pontil Arthur Gretton (2025) 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-2512-00919.txt.

Description

Dimitri Meunier and a large team improve causal estimation in a 2025 statistics paper. Researchers often want the effect of a treatment when hidden confounders blur the data. Instrumental variable regression uses an outside lever, the instrument, to isolate the treatment's true effect without seeing every confounder.

A known approach learns spectral features, directions that capture the strongest link between treatment and instrument. Those features ignore the outcome variable, which wastes information. The team's outcome-aware spectral features use the outcome too. Sharper features could give more accurate effect estimates in health and economics studies.

Catalog record

CREATORDimitri Meunier [email protected], Jakub Wornbard [email protected], Vladimir R. Kostic [email protected] Antoine MoulinAlek Fröhlich Karim Lounici, Massimiliano Pontil Arthur Gretton
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE119 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-00919

Browse the full library · About the Tower of Babel