cogintro

Cogito NTNU's Introduction Group

Overview

This repository contains materials and projects from Cogito NTNU's introduction group. Unlike regular Cogito projects that focus on a single target throughout the semester, CogIntro covers a broad range of AI/ML topics to maximize learning. The group of 8 members met twice a week in afternoon sessions, with each subproject spanning 1-4 weeks. The progression went from foundational tooling through increasingly complex applications, ending with a deep learning project on HPC infrastructure.

Course Progression

flowchart LR
    A[Project Start] --> B[Git]
    B --> C[ML Fundamentals]
    C --> D[Flappy Bird RL]
    D --> E[LLM Chatbot]
    E --> F[Tumor Segmentation]
    F --> G[Project Presentation]

    style A fill:#374151,stroke:#6b7280
    style G fill:#374151,stroke:#6b7280

Loading

Projects

Fundamentals

Terminal usage, Git workflows, and introductory ML concepts via Kaggle notebooks.

Flappy Bird Agent

A reinforcement learning agent trained to play Flappy Bird using the OpenAI Gym framework.

Project folder →

LLM Chatbot

A conversational chatbot built with OpenAI's Responses API.

Project folder →

Tumor Segmentation on IDUN

Medical image segmentation using U-Net architecture, trained on NTNU's IDUN HPC cluster via SLURM. The model segments tumor regions from PET/CT scans using k-fold cross-validation.

Project folder →

Prerequisites

• Git: Download Git
• Python 3.12: Download Python
• UV: Python package manager. Install UV

Getting Started

git clone https://github.com/CogitoNTNU/cogintro.git
cd cogintro
uv sync

For development:

uv run pre-commit install

Repository Structure

cogintro/
├── src/
│   ├── flappy-bird-gym/     # RL environment and agent
│   ├── large-language-models/  # OpenAI chatbot notebook
│   └── tumor-segmentation/  # U-Net model and SLURM scripts
├── docs/
└── tests/

Team

maiahi	AlMinaDO	arlindakm	flatval
Knolldus	Jarandvs	ApatShe	svemyh

License

Distributed under the MIT License. See LICENSE for more information.