Inspiration comes from watching participants and how much knowledge a person has reached

What it does

Based on the presence of the artist among us, the existing graphics represented the face, and therefore the image was chosen and combined with special terrain images (Google Map) to form a single panel.

How I built it

Our unique idea is based on XAI (explainable AI) as a form of art. Things like SHAP gradient explainer, visualising Activation layers, occlusion sensitivity, Grad-CAM and SmoothGrad. These technologies while important in testing ai algorithms, our great idea is to use them as a form of art and the trick is to allow the artists to choose the images and data that will be fed to these training models.

Challenges I ran into

Most of our technical difficulties require that we need time to get the job done correctly. The floydhub shutdown midway during the training. Therefore we regret moving from AWS to floydhub.

Accomplishments that I'm proud of

XAIArt Framework based on explainable AI as a form of art. In addition, the ability to teach and code up gan from scratch and training it with artists choosing the data.

What I learned

This wonderful environment is a fertile place for more knowledge, and teamwork increases knowledge and visual nutrition.

What's next for Human Civilization_112

Prepare panels of different sizes and arrange them in the form of buildings to communicate the idea to the world

Share this project: