Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Awakening Dormant Users: Generative Recommendation with Counterfactual Functional Role Reasoning

Huishi Luo, Shuokai Li, Hanchen Yang, Zhongbo Sun, Haojie Ding, Boheng Zhang, Zijia Cai, Renliang Qian, Fan Yang, Tingting Gao, Chenyi Lei, Wenwu Ou, Fuzhen Zhuang

TL;DR

RoleGen, a novel framework that synergizes a Conversion Trajectory Reasoner with a Generative Behavioral Backbone, explicitly models the context-dependent Functional Role of items to reconstruct intent evolution and employs counterfactual inference to simulate diverse conversion paths, effectively mitigating interest collapse.

Abstract

Awakening dormant users, who remain engaged but exhibit low conversion, is a pivotal driver for incremental GMV growth in large-scale e-commerce platforms. However, existing approaches often yield suboptimal results since they typically rely on single-step estimation of an item's intrinsic value (e.g., immediate click probability). This mechanism overlooks the instrumental effect of items, where specific interactions act as triggers to shape latent intent and drive subsequent decisions along a conversion trajectory. To bridge this gap, we propose RoleGen, a novel framework that synergizes a Conversion Trajectory Reasoner with a Generative Behavioral Backbone. Specifically, the LLM-based Reasoner explicitly models the context-dependent Functional Role of items to reconstruct intent evolution. It further employs counterfactual inference to simulate diverse conversion paths, effectively mitigating interest collapse. These reasoned candidate items are integrated into the generative backbone, which is optimized via a collaborative "Reasoning-Execution-Feedback-Reflection" closed-loop strategy to ensure grounded execution. Extensive offline experiments and online A/B testing on the Kuaishou e-commerce platform demonstrate that RoleGen achieves a 6.2% gain in Recall@1 and a 7.3% increase in online order volume, confirming its effectiveness in activating the dormant user base.

Awakening Dormant Users: Generative Recommendation with Counterfactual Functional Role Reasoning

TL;DR

RoleGen, a novel framework that synergizes a Conversion Trajectory Reasoner with a Generative Behavioral Backbone, explicitly models the context-dependent Functional Role of items to reconstruct intent evolution and employs counterfactual inference to simulate diverse conversion paths, effectively mitigating interest collapse.

Abstract

Awakening dormant users, who remain engaged but exhibit low conversion, is a pivotal driver for incremental GMV growth in large-scale e-commerce platforms. However, existing approaches often yield suboptimal results since they typically rely on single-step estimation of an item's intrinsic value (e.g., immediate click probability). This mechanism overlooks the instrumental effect of items, where specific interactions act as triggers to shape latent intent and drive subsequent decisions along a conversion trajectory. To bridge this gap, we propose RoleGen, a novel framework that synergizes a Conversion Trajectory Reasoner with a Generative Behavioral Backbone. Specifically, the LLM-based Reasoner explicitly models the context-dependent Functional Role of items to reconstruct intent evolution. It further employs counterfactual inference to simulate diverse conversion paths, effectively mitigating interest collapse. These reasoned candidate items are integrated into the generative backbone, which is optimized via a collaborative "Reasoning-Execution-Feedback-Reflection" closed-loop strategy to ensure grounded execution. Extensive offline experiments and online A/B testing on the Kuaishou e-commerce platform demonstrate that RoleGen achieves a 6.2% gain in Recall@1 and a 7.3% increase in online order volume, confirming its effectiveness in activating the dormant user base.
Paper Structure (40 sections, 8 equations, 6 figures, 9 tables)

This paper contains 40 sections, 8 equations, 6 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Method
  4. Problem Formulation
  5. Overview
  6. Item Semantic-ID Alignment
  7. Functional Role Modeling
  8. Motivation.
  9. Functional Role Definition
  10. Role Trajectory Construction.
  11. Instruction Tuning with Structured Reasoning
  12. Counterfactual Functional Role Inference
  13. Motivation.
  14. Solution.
  15. Generative Behavioral Backbone
  16. ...and 25 more sections

Figures (6)

  • Figure 1: Instrumental effect in conversion trajectories. An item's utility extends beyond its immediate conversion probability. Instead, its contribution to the final conversion dynamically depends on the context and user intent.
  • Figure 2: The overall architecture of RoleGen. The framework consists of two synergistic modules: (1) The Conversion Trajectory Reasoner (top), which infers the latent Functional Role Trajectory from sparse user behaviors and employs Counterfactual Inference to explore diverse conversion paths; and (2) The Generative Behavioral Executor (bottom), which grounds these reasoning signals into precise item retrieval via a generative backbone. The two modules are optimized through a collaborative "reasoning–execution–feedback–reflection" closed-loop training strategy.
  • Figure 3: Framework of Online Deployment of RoleGen.
  • Figure 4: Ablation study of RoleGen.
  • Figure 5: Analysis of Matthew effect mitigation. The exposure ratio rises significantly above 1.0 for long-tail items.
  • ...and 1 more figures

Theorems & Definitions (2)

  • definition 1: Dormant Users
  • definition 2: Functional Role Trajectory