Deep-learning based Personalized Ads Generation

Deep-learning based Personalized Ads Generation

Deep-learning based Personalized Ads Generation

Industry Challenge/ Challenges

The client wanted to analyze the existing ads and videos to discover new patterns of customer preference and automate the content to generate more targeted advertisements. By analyzing the current textual and visual factors like images, videos, body text, and headline, the client wanted to:

  • Keywords
    To deliver better analytics, a method to auto-generate ad texts with keywords was desired.
  • Speed
    The client was looking for a faster ad-generation system with automated suggestions.
  • Cost
    A solution to improve the cost metrics prediction like CPR, CPC, CTR, etc, and cut down on costs by eliminating the need for writing content.

Problem Statement

  • Create a solution that would save marketers time and money by generating personalized ad copy and media, which could be launched straight to social media.

Our Solution

  • To analyze the ads and obtain the desired insights, we used the Shapley Additive explanations (SHAP) model.
  • We employed the Random Forest Regressor to forecast the cost metrics for various images, videos, body text, and headlines.
  • In order to generate unique text that could be used by the client for the ads, we used Generative Adversarial Network (GAN)
  • The solution enabled the platform to generate keyword-enabled text for the advertisement body and headline, which not only identified important features and patterns but also provided better metrics for targeting the ads.¬†

Tech Stack

  • Language
    • Python
  • Web Development
    • PHP, HTML, flask, CSS, bootstrap


  • Random forest
  • SHAP (for diagnostic model)
  • GPT-3 (for generating ads description)¬†
  • Google Cloud
    • Google Vision API (for extracting entities from image)
    • Google NLP API (for extracting entities from text from ads description)

Business Impact

  • We were able to scale the company by driving more conversions for their ads at lower costs, resulting in higher profits.¬†
  • With the assistance of AI, the ads were tailored to each audience segment in order to reach a large number of people.
  • Our client observed up to a 3x boost in the ROAS.

Deep-learning based Personalized Ads Generation


Recent Posts

Please fill in the form below

Subscribe to Newsletter

Stay updated with the blogs by subscribing to the newsletter