Multi-Treatment Multi-Task Uplift Modeling for Enhancing User Growth

TL;DR

MTMT addresses the challenge of multi-treatment multi-task uplift by explicitly modeling a base treatment uplift $\hat{\bar{\tau}}^k(x)$ and incremental uplifts $\bar{\tau}^k_m(x)$ using a MMOE-based user encoder and a self-attention treatment interaction module. The framework combines base and incremental uplifts to predict outcomes $\bar{y}_i^k(m)$, optimized via a joint $L_2$ loss over control and treatment groups for end-to-end estimation and deployment. Empirical results on the CRITEO and product datasets show MTMT outperforms strong baselines in both single- and multi-task, as well as multi-treatment scenarios, and the model has been deployed online in a gaming platform. The work offers a scalable, interpretable approach for personalized incentives with explicit disentanglement of base versus treatment-specific effects, enabling more precise decisions in dynamic environments.

Abstract

As a key component in boosting online user growth, uplift modeling aims to measure individual user responses (e.g., whether to play the game) to various treatments, such as gaming bonuses, thereby enhancing business outcomes. However, previous research typically considers a single-task, single-treatment setting, where only one treatment exists and the overall treatment effect is measured by a single type of user response. In this paper, we propose a Multi-Treatment Multi-Task (MTMT) uplift network to estimate treatment effects in a multi-task scenario. We identify the multi-treatment problem as a causal inference problem with a tiered response, comprising a base effect (from offering a treatment) and an incremental effect (from offering a specific type of treatment), where the base effect can be numerically much larger than the incremental effect. Specifically, MTMT separately encodes user features and treatments. The user feature encoder uses a multi-gate mixture of experts (MMOE) network to encode relevant user features, explicitly learning inter-task relations. The resultant embeddings are used to measure natural responses per task. Furthermore, we introduce a treatment-user feature interaction module to model correlations between each treatment and user feature. Consequently, we separately measure the base and incremental treatment effect for each task based on the produced treatment-aware representations. Experimental results based on an offline public dataset and an online proprietary dataset demonstrate the effectiveness of MTMT in single/multi-treatment and single/multi-task settings. Additionally, MTMT has been deployed in our gaming platform to improve user experience.

Figures (5)

• Figure 1: Activity growth rates by base treatment and multi-treatment deployment on our online gaming server. Growth rates are shown separately based on users' historical activity, as past activity significantly impacts future activity.
• Figure 2: Illustration of the proposed Multi-treatment Multi-task (MTMT) framework. Note that for clarity we only show the network structure of two tasks.
• Figure 3: The proposed user-treatment feature interaction module
• Figure 4: Distributions of base and incremental treatment effects. The base treatment effects are numerically much larger than the incremental treatment effects.
• Figure 5: Overview of our online bonus deployment platform