Counterfactual Generative Models for Time-Varying Treatments

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wu, Shenghao, Zhou, Wenbin, Chen, Minshuo, Zhu, Shixiang
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911953940643840
author Wu, Shenghao
Zhou, Wenbin
Chen, Minshuo
Zhu, Shixiang
author_facet Wu, Shenghao
Zhou, Wenbin
Chen, Minshuo
Zhu, Shixiang
contents Estimating the counterfactual outcome of treatment is essential for decision-making in public health and clinical science, among others. Often, treatments are administered in a sequential, time-varying manner, leading to an exponentially increased number of possible counterfactual outcomes. Furthermore, in modern applications, the outcomes are high-dimensional and conventional average treatment effect estimation fails to capture disparities in individuals. To tackle these challenges, we propose a novel conditional generative framework capable of producing counterfactual samples under time-varying treatment, without the need for explicit density estimation. Our method carefully addresses the distribution mismatch between the observed and counterfactual distributions via a loss function based on inverse probability re-weighting, and supports integration with state-of-the-art conditional generative models such as the guided diffusion and conditional variational autoencoder. We present a thorough evaluation of our method using both synthetic and real-world data. Our results demonstrate that our method is capable of generating high-quality counterfactual samples and outperforms the state-of-the-art baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2305_15742
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Counterfactual Generative Models for Time-Varying Treatments
Wu, Shenghao
Zhou, Wenbin
Chen, Minshuo
Zhu, Shixiang
Machine Learning
Methodology
Estimating the counterfactual outcome of treatment is essential for decision-making in public health and clinical science, among others. Often, treatments are administered in a sequential, time-varying manner, leading to an exponentially increased number of possible counterfactual outcomes. Furthermore, in modern applications, the outcomes are high-dimensional and conventional average treatment effect estimation fails to capture disparities in individuals. To tackle these challenges, we propose a novel conditional generative framework capable of producing counterfactual samples under time-varying treatment, without the need for explicit density estimation. Our method carefully addresses the distribution mismatch between the observed and counterfactual distributions via a loss function based on inverse probability re-weighting, and supports integration with state-of-the-art conditional generative models such as the guided diffusion and conditional variational autoencoder. We present a thorough evaluation of our method using both synthetic and real-world data. Our results demonstrate that our method is capable of generating high-quality counterfactual samples and outperforms the state-of-the-art baselines.
title Counterfactual Generative Models for Time-Varying Treatments
topic Machine Learning
Methodology
url https://arxiv.org/abs/2305.15742