joshuaknauber / cs50

CS50 Course - Joshua Knauber

This repository contains my solutions to some of the exercises in the CS50 course. It also contains some additional projects around similar topics.

Setup

To run the files first create a virtual environment and install the requirements:

python3 -m venv venv
source venv/bin/activate

pip install -r requirements.txt

Classification

Shopping

From CS50, based on the given csv with shopping data train a KNeighboursClassifier to predict whether a customer will buy a product or not.

The functions to implement were load_data, train_model and evaluate.

Genetic

Function

A project to predict a given function using a genetic algorithm.

Knapsack

A project to solve the knapsack problem using a genetic algorithm.

Knowledge

Minesweeper

From CS50, run runner.py to play. The functions add_knowledge, make_safe_move and make_random_move were left to implement.

Neural Nets

MNIST

A project to predict numbers from the mnist dataset using a tensorflow. Experiment notes are in the notebook.

Traffic Signs

From CS50, project to predict categories of traffic signs from the GTSRB dataset. More notes in the README in the folder.

Optimization

Crossword

From CS50, project to generate a crossword puzzle with backtracking. The functions to implement were enforce_node_consistency, revise, ac3, assignment_complete, consistent, order_domain_values, select_unassigned_variable and backtrack.

Sudoku

A project to solve Sudoku with backtracking, implementation under search/sudoku.

Reinforcement Learning

Nim

From CS50, project to play Nim against an AI. The functions to implement were get_q_value, update_q_value, best_future_reward and choose_action.

Search

Maze

A project to implement different search algorithms from scratch. There are different frontier classes that can be exchanged to use them.

Sudoku

An attempt to use search algorithms to solve Sudoku. The backtracking implementation is also here.

Tic Tac Toe

A project to play tic tac toe with minimax players. There is one implementation with and without alpha beta pruning.