Dynamic Product Image Generation and Recommendation at Scale for Personalized E-commerce

TL;DR

This paper presents a scalable method combining latent diffusion image generation and contextual bandits to create personalized product images, significantly enhancing user engagement in e-commerce retargeting campaigns.

Contribution

It introduces a novel integration of diffusion-based image synthesis with contextual bandits for scalable personalized product image generation.

Findings

Increased user engagement in retargeting campaigns

Effective scalable personalized image creation

Demonstrated practical application in e-commerce

Abstract

Coupling latent diffusion based image generation with contextual bandits enables the creation of eye-catching personalized product images at scale that was previously either impossible or too expensive. In this paper we showcase how we utilized these technologies to increase user engagement with recommendations in online retargeting campaigns for e-commerce.