Cross-Head Attention Uplift Network with Inverse Propensity Score under Unobserved Confounding
For practitioners in causal inference and e-commerce, this work provides a method to better estimate individual treatment effects when unobserved confounders exist, but the improvements are incremental over existing deep learning approaches.
The paper proposes the Cross-Head Attention Uplift Network (CHAUN) and Robust Adversarial Inverse Propensity Score (RA-IPS) to improve uplift modeling under unobserved confounding. Experiments show CHAUN achieves up to 25.6% relative improvement in QINI scores over state-of-the-art models, and RA-IPS outperforms standard IPS by 5.4%.
Uplift modeling, crucial for estimating individual treatment effects (ITE), faces dual challenges: flexibly leveraging inter-group similarity to enhance discriminative power and debiasing under unobserved confounding scenarios. In this paper, we propose the Cross-Head Attention Uplift Network (CHAUN) and Robust Adversarial Inverse Propensity Score (RA-IPS) method to address these limitations. CHAUN employs shared feature embeddings and cross-head attention mechanisms to dynamically integrate treatment-specific and control-specific representations, enhancing inter-group correlation modeling. Theoretically, we prove that access to the true propensity scores ensures ITE identifiability even with unobserved confounders. For practical scenarios lacking true propensity scores, RA-IPS adversarially optimizes propensity weights within constrained uncertainty sets to mitigate bias from unobserved variables. Experiments on public datasets (CRITEO-UPLIFT, LAZADA) and a production e-commerce dataset demonstrate CHAUN's superiority over state-of-the-art uplift models, achieving relative improvements of up to 25.6% in QINI scores. RA-IPS further enhances robustness, outperforming standard IPS by 5.4% under unobserved confounding. The results validate the effectiveness of our proposed methods in real-world causal inference tasks.