The Ad Machine
This work is a collaboration between Yale University and Shenkar College of Engineering and Design. Using research from Yale University - “Uninformative Advertising as an Invitation to Search” - we developed a research that explores the impact of design on a modern-day, globalized, informed consumer. We build on the notion of uninformed advertising where viewers receive less information which in turn piques their curiosity and results in heightened engagement. We take the concept one step further and ask what would happen if we gave a computer the task of creating nonlogical advertisements baised Machine learing models will people find the real from the fake? What happens when you take the rules of traditional advertising and design creations process and try pushing their logical boundirs? Will machines one day learn to be creative? In the absence logic or mining does the medium still have the same intended impact?
First step
We collected 48 Magazines And their ads in order to decode their CSS
Second step - Machine learning
We used KNN to Cluster the data and find new designs based on the clusters the centroid
Third step - Coding the CSS of the generated grades
We coded the grades in order to start generating ads based on our generative design model
Third step - Using Open API's
To inject content images and text to the grids to formed the ad machine
The Script
Here is an example using Real Ad's from the New Yourk Times
We stripped the ads out of their content to unveil their grid system
the Ad's grids system decoded
Once the grid is unveiled we code it
Random generated ads using CSS and API's
After having the CSS of the advertisement we can integrate it to Content API's for photos text and logos
Finding the real ad out of the fake computer generated ads it's not as easy as it looks
Can you spot out the fake ones?
