My Completed Projects from Udacity's Deep Learning Nanodegree

Author: Eremey Valetov
Course completion date: 18-May-2018
Email: valetove@msu.edu

This repository contains the projects I have completed in Udacity's Deep Learning Nanogree in Spring 2018.

Project 1: Introductory Project — Predicting Bike Sharing Data

Directory: project_1_first_neural_network
Description: a fully connected neural network is used to predict future bike-sharing service usage based on past data.

Project 2: Convolutional Neural Networks — Dog Breed Classifier

Directory: project_2_CNN_dog_breed_classifier
Description: a Convolutional Neural Network (CNN) is used to determine whether an image contains a human, a dog, or neither. If a dog is detected in the image, the program attempts to guess the breed. If a human is detected, the program returns the resembling dog breed.

Project 3: Recurrent Neural Networks — Generate TV Scripts

Directory: project_3_RNN_TV_script_generation
Description: a Recurrent Neural Network (RNN) is used to generate text resembling a TV script based on past episodes of the Simpsons.

Project 4: Generative Adversarial Networks — Generate Faces

Directory: project_4_GAN_face_generation
Description: a Generative Adversarial Network (GAN) is used to generate images resembling human faces based on a dataset of celebrity images.

Project 5: Deep Reinforcement Learning — Teach a Quadcopter to Fly

Directory: project_5_RL_quadcopter_project
Description: a Deep Deterministic Policy Gradient (DDPG) method is used to train a model quadcopter to take off and reach the target altitude.