Disentangled Representation Learning for Causal Inference With Instruments

doi:10.1109/TNNLS.2024.3512790

Disentangled Representation Learning for Causal Inference With Instruments

2024-01-01

0

PRE

AI

DOI：10.1109/TNNLS.2024.3512790

原文链接

原文求助

摘要

En 中文

Latent confounders are a fundamental challenge for inferring causal effects from observational data. The instrumental variable (IV) approach is a practical way to address this challenge. Existing IV-based estimators need a known IV or other strong assumptions, such as the existence of two or more IVs in the system, which limits the application of the IV approach. In this article, we consider a relaxed requirement, which assumes there is an IV proxy in the system without knowing which variable is the proxy. We propose a variational autoencoder (VAE)-based disentangled representation learning method to learn an IV representation from a dataset with latent confounders and then utilize the IV representation to obtain an unbiased estimation of the causal effect from the data. Extensive experiments on synthetic and real-world data have demonstrated that the proposed algorithm outperforms the existing IV-based estimators and VAE-based estimators.

Keyword:

Instruments

Estimation

Mathematical models

Disentangled representation learning

Australia

Object recognition

Learning systems

Computational modeling

Training

Time factors

Causal inference

disentangled representation learning

instrumental variable (IV)

latent variables

observational data

Disentangled Representation Learning for Causal Inference With Instruments

摘要

期刊

机构

引用论文