Evaluating Deep Vs. Wide & Deep Learners As Contextual Bandits For Personalized Email Promo Recommendations

TL;DR

This paper compares wide and deep neural networks for personalized email promotion recommendations modeled as a contextual bandit problem, showing similar accuracy and slight improvements with Bayesian methods.

Contribution

It demonstrates that wide and deep models perform comparably in predicting promotional offers and explores the impact of categorical features and Bayesian sampling methods.

Findings

Similar prediction accuracy for wide and deep models.

Including categorical features improves accuracy depending on feature variability.

Bayesian methods like Monte Carlo dropout slightly enhance model performance.

Abstract

Personalization enables businesses to learn customer preferences from past interactions and thus to target individual customers with more relevant content. We consider the problem of predicting the optimal promotional offer for a given customer out of several options as a contextual bandit problem. Identifying information for the customer and/or the campaign can be used to deduce unknown customer/campaign features that improve optimal offer prediction. Using a generated synthetic email promo dataset, we demonstrate similar prediction accuracies for (a) a wide and deep network that takes identifying information (or other categorical features) as input to the wide part and (b) a deep-only neural network that includes embeddings of categorical features in the input. Improvements in accuracy from including categorical features depends on the variability of the unknown numerical features for each category. We also show that selecting options using upper confidence bound or Thompson sampling, approximated via Monte Carlo dropout layers in the wide and deep models, slightly improves model performance.