NTUA AI Coursework

Student: Michael-Athanasios Peppas (03121026) Institution: National Technical University of Athens, ECE

This repository includes Google Colab notebooks from four NTUA ECE courses: Artificial Intelligence, Machine Learning, Neural Networks & Deep Learning, and Image and Video Technology and Analysis. Each course contains two comprehensive lab projects with supporting code, data, references, and experiment notes.

---

Courses

Artificial Intelligence

• AI_Lab1: Maze generation, search algorithms (BFS, Dijkstra, A*, Greedy), and adversarial planning with Alpha-Beta agents.
• AI_Lab2: Hybrid movie recommender combining SWI-Prolog symbolic reasoning and Python for data handling and evaluation.

See Artificial_Intelligence/README.md for detailed instructions, key concepts, and results.

Machine Learning

• ML_Lab1: RainTomorrow classification: data preprocessing, feature engineering, model training with scikit-learn, and evaluation.
• ML_Lab2: Salinas hyperspectral clustering and classification using KMeans, Fuzzy C-Means, PCA, and CNN feature extraction.

See Machine_Learning/README.md for detailed instructions, key concepts, and results.

Neural Networks & Deep Learning

• DL_Lab1: WideResNet architectures on CIFAR-10, Mixup augmentation, calibration measurement, and robustness on CIFAR-10-C.
• DL_Lab2: Transformer fine-tuning for sentiment, PiQA, TruthfulQA, and Winogrande tasks using Hugging Face.

See Neural_Networks_and_Deep_Learning/README.md for detailed instructions, key concepts, and results.

Image and Video Technology and Analysis

• Lab 1: Laplacian pyramid image coding with Gaussian/Laplacian pyramid construction, exact reconstruction, entropy analysis, and quantization experiments.
• Lab 2: CNN image classification on a 20-class CIFAR-100 subset using LeNet, AlexNet, VGG, MyCNN, dropout/data augmentation, VGG19, and EfficientNetB0.

See Image_and_Video_Technology_and_Analysis/README.md for detailed instructions, key concepts, and results.

---

Repository Structure

ntua-ai-coursework/
├── Artificial_Intelligence/
│   ├── AI_Lab1/
│   ├── AI_Lab2/
│   └── README.md
├── Machine_Learning/
│   ├── ML_Lab1/
│   ├── ML_Lab2/
│   └── README.md
├── Neural_Networks_and_Deep_Learning/
│   ├── DL_Lab1/
│   ├── DL_Lab2/
│   └── README.md
├── Image_and_Video_Technology_and_Analysis/
│   ├── lab1/
│   ├── lab2/
│   ├── lab_material/
│   └── README.md
├── docs/
│   └── project_structure/
├── LICENSE
└── README.md

---

Prerequisites

• Python 3.8+

• Core Python libraries:

  pip install numpy pandas matplotlib scikit-learn scikit-fuzzy pillow scikit-image tensorflow torch torchvision torchaudio transformers datasets evaluate sentence-transformers tqdm jupyter

• SWI-Prolog (v8.x): required for AI_Lab2

  sudo apt-get install swi-prolog
  pip install pyswip

Some notebooks download public datasets at runtime. For DL_Lab1, download and extract CIFAR-10-C into the path specified by the course README.

---

Getting Started

1. Clone the repository

   git clone <repo_url> ntua-ai-coursework
   cd ntua-ai-coursework

2. Install dependencies as listed above.

3. Explore each course by opening its README.md and running the lab notebooks in Google Colab or Jupyter.

4. Keep lab support folders beside their notebooks so local imports, data files, PDFs, and saved experiment artifacts resolve correctly.

---

License

This project is licensed under the MIT License. See LICENSE for details.

---

Prepared by Michael-Athanasios Peppas (03121026)