HUPLACLIP-NNs

This repository evaluates the performance of various AI models on the Planted Clique Problem, a theoretical computer science task focused on detecting hidden cliques within graphs.

Repository Structure

• main.py: Main script for running experiments.
• README.md: Project documentation.
• requirements.txt: Lists the necessary Python packages and dependencies on personal laptop.
• requirements_Leo.txt: Lists the necessary Python packages and dependencies Leonardo HPC cluster.
• .gitignore: Specifies files and directories to be ignored by Git.

Folders

• docs/: Contains documentation files.
  • grid_exp_config.yml: Specifies all experiment parameters, such as model types, hyperparameters, and graph settings.
• results/: Directory where experiment results are saved.
• scripts/: Additional scripts for setup, preprocessing, or visualizations.
• src/: Source files that include modules for data generation, model definitions, and training functions.
• tests/: Contains unit tests for the various functions and modules in the project.

Models

The AI models tested include:

• MLP (Multi-Layer Perceptron)
• CNN (Convolutional Neural Network)
• ViT (Vision Transformer)

Contact

For inquiries, please contact:
Daniele Tirinnanzi - daniele.tirinnanzi@unicampus.it