Hybrid Model-Based RL for Intelligent Marketing: Dyna-Q Meets Transformer Models and Bayesian Survival Priors

Author(s): Shenggang Li

Originally published on Towards AI.

A theory-to-practice study on profit-driven customer re-engagement in e-commerce using BG/NBD-augmented Attention and budget-aware roll-outs

We built a next-gen coupon engine fusing three techniques: a Bayesian survival model for repurchase chance, an attention-based Transformer for profit forecasting, and a Dyna-Q RL agent for continuous policy optimization — forming a self-optimizing loop.

The article discusses the implementation of a coupon engine that integrates a Bayesian model for predicting customer returns, an attention-based Transformer to forecast profits, and a Dyna-Q reinforcement learning agent for policy optimization. Focusing on profit-driven customer re-engagement in e-commerce, it highlights the benefits of filtering customers based on their likelihood to return, predicting profit from recent behaviors, and using simulations to optimize coupon strategies. The system effectively combines these models to minimize testing costs while maximizing returns, illustrating its capability to adapt to customer behaviors and improve marketing effectiveness.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI