How shahriar-0/artificial-intelligence-course-projects-s2023 works
Main Components
• Course Projects Container – This is the overall “wrapper” for the repository.
• Individual Project Modules – Represent each numbered assignment:
• CA0: Introduction (Exploratory Data Analysis using Pandas, NumPy, and Matplotlib)
• CA1: Search Algorithms (Implementing search methods like BFS, IDS, and A* for a recipes graph problem)
• CA2: Genetic Algorithms (Investment decision making via genetic optimization)
• CA3: Game Minimax (Game playing using the minimax algorithm with alpha-beta pruning, tested on an Othello game)
• CA4: Naive Bayes (Different Naïve Bayes techniques for classification)
• CA5: Machine Learning (Multiple traditional ML models on a housing dataset with feature engineering and model evaluation)
• CA6: Neural Networks (Divided into two phases – Phase 1 using custom feedforward networks and PyTorch; Phase 2 using Keras/TensorFlow for CIFAR-10 classification)
• Supporting Materials – These include additional Python demo notebooks (for Pandas, Matplotlib, NumPy) and tutorial content outside of the course projects.
• Archives – Older versions or experimental builds (“Archive”, “Genetic-Archive”) that might indicate versioning or historical snapshots.
Relationships and Interactions
• Each module is essentially self-contained and includes its own Jupyter Notebook(s), asset subfolders (for data and introductory/description PDFs), test files, and supporting datasets. They do not interdepend on one another—each project functions independently.
• The Python directory (with “Demo_Materials”, “Matplotlib”, “Numpy”, “Pandas”) is a separate resource that provides learning examples or additional tutorial material for users to experiment with core Python data science libraries.
• Some modules have further sub-structure (e.g., CA6 has two phases) which should be depicted as nested components within the Neural Networks module.
• For each project, the flow is essentially from:
Input Data (CSV, PKL, etc.) → Algorithm/Notebook Implementation (e.g., search, minimax, ML model training) → Output (results stored or printed, sometimes described in asset files).
• The Assets directories show additional resources (e.g., PDFs for project description) that can be noted as “Documentation/Supporting Information” for that module.
Architectural Patterns or Design Principles
• Modular Design – Each assignment is a modular component that encapsulates a distinct AI algorithm or learning exercise.
• Separation of Concerns – The data exposition, algorithm logic, evaluation outputs, and documentation are well separated into individual subfolders within each module.
• Self-Contained Experimentation – Each module contains its own notebooks and datasets, enabling isolated testing and evaluation.
• Versioning/Archival – The “Archive” directories suggest iterative improvements or historical snapshots that can be depicted as a timeline or backup branch for reference.