Cross-Treatment Effect Estimation for Multi-Category, Multi-Valued Causal Inference via Dynamic Neural Masking

TL;DR

This paper introduces XTNet, a neural network architecture that effectively estimates effects of complex multi-category, multi-valued treatments, capturing interactions without restrictive assumptions, and validated through extensive experiments and real-world testing.

Contribution

The paper presents XTNet with a novel cross-effect estimation module and dynamic masking, enabling scalable and accurate modeling of multi-valued treatment interactions in causal inference.

Findings

XTNet outperforms existing methods in ranking accuracy.

It achieves superior effect estimation quality.

Real-world A/B tests confirm its practical effectiveness.

Abstract

Counterfactual causal inference faces significant challenges when extended to multi-category, multi-valued treatments, where complex cross-effects between heterogeneous interventions are difficult to model. Existing methodologies remain constrained to binary or single-type treatments and suffer from restrictive assumptions, limited scalability, and inadequate evaluation frameworks for complex intervention scenarios. We present XTNet, a novel network architecture for multi-category, multi-valued treatment effect estimation. Our approach introduces a cross-effect estimation module with dynamic masking mechanisms to capture treatment interactions without restrictive structural assumptions. The architecture employs a decomposition strategy separating basic effects from cross-treatment interactions, enabling efficient modeling of combinatorial treatment spaces. We also propose MCMV-AUCC, a suitable evaluation metric that accounts for treatment costs and interaction effects. Extensive experiments on synthetic and real-world datasets demonstrate that XTNet consistently outperforms state-of-the-art baselines in both ranking accuracy and effect estimation quality. The results of the real-world A/B test further confirm its effectiveness.