ArtIfact: Detecting AI-generated Art

Videos

• Platform Demo: https://youtu.be/CMAa6CjUErY
• Theoretical Background: https://youtu.be/kfeess6JJWk

backend

The backend processes uploaded images and responds with model predictions. Built with the following technologies:

• AWS Lambda: Request handling, routing to other AWS services
• AWS S3: Store uploaded images and model artifacts
• AWS RDS: Store in-progress jobs and image metadata
• AWS SageMaker: Deploy model for serverless inference
• AWS ECR: Store Docker images containing model inference code

client

• The client CLI is used to upload images for analysis and retrieve generated predictions.
• Running client.py requires the requests library, which can be installed with:

pip install requests

ml

• Contains model training notebooks and saved model parameters.
• Machine learning was conducted on the PyTorch deep learning platform.
• Read SETUP.md for instructions for installing dependencies