Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A New Creative Generation Pipeline for Click-Through Rate with Stable Diffusion Model

Hao Yang, Jianxin Yuan, Shuai Yang, Linhe Xu, Shuo Yuan, Yifan Zeng

TL;DR

The paper addresses CTR optimization for online ad creatives by tightly integrating generation with optimization targets. It introduces CG4CTR, a diffusion-based pipeline that inpaints background content while preserving the product, and uses a Prompt Model for personalized prompts and a multimodal Reward Model to rank and train the generators, all within a self-cycling training loop. Quantitative online results show meaningful CTR and revenue gains across multiple categories, while ablations demonstrate the value of user-aware prompts, multimodal fusion, pre-training, and iterative refinement. The approach offers practical impact by enabling CTR-driven, personalized creative generation that can be integrated into existing recommendation pipelines to improve online advertising performance.

Abstract

In online advertising scenario, sellers often create multiple creatives to provide comprehensive demonstrations, making it essential to present the most appealing design to maximize the Click-Through Rate (CTR). However, sellers generally struggle to consider users preferences for creative design, leading to the relatively lower aesthetics and quantities compared to Artificial Intelligence (AI)-based approaches. Traditional AI-based approaches still face the same problem of not considering user information while having limited aesthetic knowledge from designers. In fact that fusing the user information, the generated creatives can be more attractive because different users may have different preferences. To optimize the results, the generated creatives in traditional methods are then ranked by another module named creative ranking model. The ranking model can predict the CTR score for each creative considering user features. However, the two above stages are regarded as two different tasks and are optimized separately. In this paper, we proposed a new automated Creative Generation pipeline for Click-Through Rate (CG4CTR) with the goal of improving CTR during the creative generation stage. Our contributions have 4 parts: 1) The inpainting mode in stable diffusion is firstly applied to creative generation task in online advertising scene. A self-cyclic generation pipeline is proposed to ensure the convergence of training. 2) Prompt model is designed to generate individualized creatives for different user groups, which can further improve the diversity and quality. 3) Reward model comprehensively considers the multimodal features of image and text to improve the effectiveness of creative ranking task, and it is also critical in self-cyclic pipeline. 4) The significant benefits obtained in online and offline experiments verify the significance of our proposed method.

A New Creative Generation Pipeline for Click-Through Rate with Stable Diffusion Model

TL;DR

The paper addresses CTR optimization for online ad creatives by tightly integrating generation with optimization targets. It introduces CG4CTR, a diffusion-based pipeline that inpaints background content while preserving the product, and uses a Prompt Model for personalized prompts and a multimodal Reward Model to rank and train the generators, all within a self-cycling training loop. Quantitative online results show meaningful CTR and revenue gains across multiple categories, while ablations demonstrate the value of user-aware prompts, multimodal fusion, pre-training, and iterative refinement. The approach offers practical impact by enabling CTR-driven, personalized creative generation that can be integrated into existing recommendation pipelines to improve online advertising performance.

Abstract

In online advertising scenario, sellers often create multiple creatives to provide comprehensive demonstrations, making it essential to present the most appealing design to maximize the Click-Through Rate (CTR). However, sellers generally struggle to consider users preferences for creative design, leading to the relatively lower aesthetics and quantities compared to Artificial Intelligence (AI)-based approaches. Traditional AI-based approaches still face the same problem of not considering user information while having limited aesthetic knowledge from designers. In fact that fusing the user information, the generated creatives can be more attractive because different users may have different preferences. To optimize the results, the generated creatives in traditional methods are then ranked by another module named creative ranking model. The ranking model can predict the CTR score for each creative considering user features. However, the two above stages are regarded as two different tasks and are optimized separately. In this paper, we proposed a new automated Creative Generation pipeline for Click-Through Rate (CG4CTR) with the goal of improving CTR during the creative generation stage. Our contributions have 4 parts: 1) The inpainting mode in stable diffusion is firstly applied to creative generation task in online advertising scene. A self-cyclic generation pipeline is proposed to ensure the convergence of training. 2) Prompt model is designed to generate individualized creatives for different user groups, which can further improve the diversity and quality. 3) Reward model comprehensively considers the multimodal features of image and text to improve the effectiveness of creative ranking task, and it is also critical in self-cyclic pipeline. 4) The significant benefits obtained in online and offline experiments verify the significance of our proposed method.
Paper Structure (17 sections, 21 equations, 4 figures, 3 tables, 1 algorithm)

This paper contains 17 sections, 21 equations, 4 figures, 3 tables, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORKS
  3. Text-to-image Generation
  4. Creative CTR Prediction
  5. METHODS
  6. Pipeline
  7. LoRA Model in Stable Diffusion
  8. Prompt Model
  9. Reward Model
  10. EXPERIMENTS
  11. Online Results
  12. In-depth Analysis
  13. Ablation Study on Reward Model.
  14. Ablation Study on Prompt Model.
  15. Self-cycling Analysis
  16. ...and 2 more sections

Figures (4)

  • Figure 1: Generating creatives only by modifying background using SD method in inpainting mode.
  • Figure 2: (a) Creative generation pipeline. (b) The structure of prompt model. (c) The structure of reward model.
  • Figure 3: (a) Word cloud analysis of generated tokens by prompt models without and with considering user information. (b) The creatives generated by different versions of LoRA and prompt models are scored by reward model to show the effectiveness of self-cycling training process.
  • Figure 4: Cases with the original image and generated creatives in different steps of self-cycling. Bottom score is the real CTR.