Estimating Treatment Effects under Algorithmic Interference: A Structured Neural Networks Approach

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhan, Ruohan, Han, Shichao, Hu, Yuchen, Jiang, Zhenling
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912949333917696
author Zhan, Ruohan
Han, Shichao
Hu, Yuchen
Jiang, Zhenling
author_facet Zhan, Ruohan
Han, Shichao
Hu, Yuchen
Jiang, Zhenling
contents Online user-generated content platforms allocate billions of dollars of promotional traffic through algorithms in two-sided marketplaces. To evaluate updates to these algorithms, platforms frequently rely on creator-side randomized experiments. However, because treated and control creators compete for exposure, such experiments suffer from algorithmic interference: exposure outcomes depend on competitors' treatment status. We show that commonly used difference-in-means estimators can therefore be severely biased and may even recommend deploying inferior algorithms. To address this challenge, we develop a structured semiparametric framework that explicitly models the competitive allocation mechanism underlying exposure. Our approach combines an algorithm choice model that characterizes how exposure is allocated across competing content with a viewer response model that captures engagement conditional on exposure. We construct a debiased estimator grounded in the double machine learning framework to recover the global treatment effect of platform-wide rollout. Methodologically, we extend DML asymptotic theory to accommodate correlated samples arising from overlapping consideration sets. Using Monte Carlo simulations and a large-scale field experiment on a major short-video platform, we show that our estimator closely matches an interference-free benchmark obtained from a costly double-sided experimental design. In contrast, standard estimators exhibit substantial bias and, in some cases, even reverse the sign of the effect.
format Preprint
id arxiv_https___arxiv_org_abs_2406_14380
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Estimating Treatment Effects under Algorithmic Interference: A Structured Neural Networks Approach
Zhan, Ruohan
Han, Shichao
Hu, Yuchen
Jiang, Zhenling
Econometrics
Machine Learning
Methodology
Online user-generated content platforms allocate billions of dollars of promotional traffic through algorithms in two-sided marketplaces. To evaluate updates to these algorithms, platforms frequently rely on creator-side randomized experiments. However, because treated and control creators compete for exposure, such experiments suffer from algorithmic interference: exposure outcomes depend on competitors' treatment status. We show that commonly used difference-in-means estimators can therefore be severely biased and may even recommend deploying inferior algorithms. To address this challenge, we develop a structured semiparametric framework that explicitly models the competitive allocation mechanism underlying exposure. Our approach combines an algorithm choice model that characterizes how exposure is allocated across competing content with a viewer response model that captures engagement conditional on exposure. We construct a debiased estimator grounded in the double machine learning framework to recover the global treatment effect of platform-wide rollout. Methodologically, we extend DML asymptotic theory to accommodate correlated samples arising from overlapping consideration sets. Using Monte Carlo simulations and a large-scale field experiment on a major short-video platform, we show that our estimator closely matches an interference-free benchmark obtained from a costly double-sided experimental design. In contrast, standard estimators exhibit substantial bias and, in some cases, even reverse the sign of the effect.
title Estimating Treatment Effects under Algorithmic Interference: A Structured Neural Networks Approach
topic Econometrics
Machine Learning
Methodology
url https://arxiv.org/abs/2406.14380