AI+Education Teaching Case: Computational Advertising Course Reform under Artificial Intelligence

It is reported that Shantou University’s School of Journalism and Communication recently shared an AI+ teaching case on “Computational Advertising Course Development under Artificial Intelligence Background.” This case is part of the China Association of Higher Education project titled “Research on Reform Pathways for Advertising Education in the Intelligent Communication Era.”

The project was launched in September 2024. Related content has been experimented in the Spring 2025 “Introduction to Computational Advertising” course, with a related paper awaiting publication.

Background and Needs

Traditional advertising courses focus on classic theory and case analysis, but students’ knowledge is significantly disconnected from data- and algorithm-driven modern advertising industry practices. To address issues such as “theory divorced from practice,” “cross-disciplinary knowledge segmentation,” “course content lagging behind technological development,” and “insufficient complex problem modeling skills,” the school has combined AI with computational advertising teaching for course reform.

Reform Content

AI Integration of Course Modules

Module One: Foundation and Intelligent Targeting

Traditional content: user profiling, contextual targeting, behavioral targeting.

AI integration: Machine learning empowers user profiling by using clustering algorithms (such as K-Means) for user segmentation and classification algorithms to predict user attributes (gender, age, interests).

Module Two: Core Bidding and Estimation Mechanisms

Traditional content: advertising auction mechanisms (such as GSP, VCG), CTR estimation introduction.

AI integration: Hands-on projects using open datasets (such as Criteo CTR dataset) allow students to implement basic CTR estimation models.

Module Three: Programmatic Creative and Generation

Traditional content: advertising creative design and A/B testing.

AI integration: Generative AI automatically generates ad images and videos, NLP technology generates multiple ad copies, and reinforcement learning selects optimal “image + copy” combinations for display based on user characteristics.

Resource Development

Integrating the computational advertising virtual simulation system built by the Advertising School of Communication University of China, and simulating a DSP bidding environment where students write bidding algorithms to compete.

Reform Goals and Achievements

Reform goals include: completing the “Introduction to Computational Advertising” course reform within 1 year, forming a promotable “AI + Computational Advertising” teaching model within 2 years, and applying for teaching reform projects and awards within 3 years.

Currently, the “Introduction to Computational Advertising” course practice reform has been implemented, and the paper “Exploring Advertising Education Pathways in the Intelligent Communication Era” has been published. Students can apply the approach to competitions such as computer simulation marketing contests and enhance their competitiveness in internships.

Innovation Breakthrough

The teaching paradigm has shifted from “looking backward” at case analysis to “looking forward” to systematic construction and optimization. Students no longer simply analyze “why a particular ad succeeded,” but learn how to build intelligent advertising systems that can automatically learn and optimize in real-time, developing systems engineering thinking.

Course content has evolved from static 4P theory to dynamic algorithm models as the core, progressing from logistic regression and factorization machines to Wide & Deep, DeepFM, DIN and other deep learning models.

Source: STU School of Journalism and Communication WeChat Official Account