Orthogonal Uplift Learning with Permutation-Invariant Representations for Combinatorial Treatments

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Su, Xinyan, Gao, Jiacan, Ma, Mingyuan, Xu, Xiao, Wan, Xinrui, Gu, Tianqi, Yu, Enyun, Guo, Jiecheng, Zhang, Zhiheng
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908848700260352
author Su, Xinyan
Gao, Jiacan
Ma, Mingyuan
Xu, Xiao
Wan, Xinrui
Gu, Tianqi
Yu, Enyun
Guo, Jiecheng
Zhang, Zhiheng
author_facet Su, Xinyan
Gao, Jiacan
Ma, Mingyuan
Xu, Xiao
Wan, Xinrui
Gu, Tianqi
Yu, Enyun
Guo, Jiecheng
Zhang, Zhiheng
contents We study uplift estimation for combinatorial treatments. Uplift measures the pure incremental causal effect of an intervention (e.g., sending a coupon or a marketing message) on user behavior, modeled as a conditional individual treatment effect. Many real-world interventions are combinatorial: a treatment is a policy that specifies context-dependent action distributions rather than a single atomic label. Although recent work considers structured treatments, most methods rely on categorical or opaque encodings, limiting robustness and generalization to rare or newly deployed policies. We propose an uplift estimation framework that aligns treatment representation with causal semantics. Each policy is represented by the mixture it induces over contextaction components and embedded via a permutation-invariant aggregation. This representation is integrated into an orthogonalized low-rank uplift model, extending Robinson-style decompositions to learned, vector-valued treatments. We show that the resulting estimator is expressive for policy-induced causal effects, orthogonally robust to nuisance estimation errors, and stable under small policy perturbations. Experiments on large-scale randomized platform data demonstrate improved uplift accuracy and stability in long-tailed policy regimes
format Preprint
id arxiv_https___arxiv_org_abs_2602_19851
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Orthogonal Uplift Learning with Permutation-Invariant Representations for Combinatorial Treatments
Su, Xinyan
Gao, Jiacan
Ma, Mingyuan
Xu, Xiao
Wan, Xinrui
Gu, Tianqi
Yu, Enyun
Guo, Jiecheng
Zhang, Zhiheng
Methodology
Machine Learning
We study uplift estimation for combinatorial treatments. Uplift measures the pure incremental causal effect of an intervention (e.g., sending a coupon or a marketing message) on user behavior, modeled as a conditional individual treatment effect. Many real-world interventions are combinatorial: a treatment is a policy that specifies context-dependent action distributions rather than a single atomic label. Although recent work considers structured treatments, most methods rely on categorical or opaque encodings, limiting robustness and generalization to rare or newly deployed policies. We propose an uplift estimation framework that aligns treatment representation with causal semantics. Each policy is represented by the mixture it induces over contextaction components and embedded via a permutation-invariant aggregation. This representation is integrated into an orthogonalized low-rank uplift model, extending Robinson-style decompositions to learned, vector-valued treatments. We show that the resulting estimator is expressive for policy-induced causal effects, orthogonally robust to nuisance estimation errors, and stable under small policy perturbations. Experiments on large-scale randomized platform data demonstrate improved uplift accuracy and stability in long-tailed policy regimes
title Orthogonal Uplift Learning with Permutation-Invariant Representations for Combinatorial Treatments
topic Methodology
Machine Learning
url https://arxiv.org/abs/2602.19851