LMH Summer Programme

This repository contains the projects for the tutorial during 2023 LMH Summer Programme Section 2.

Hi there, I am Jerry. You may know me if you are in the same AIML class in the summer 2023.

Colorization based on Autoencoder with CIFAR 10

This task is based on Autoencoder, which is a very basic Generating Deep Learning model. In this project, the colorization is implemented on the CIFAR 10 dataset. Where the inputs are grayscale images and the outputs are colorised ones.

Face Generator based on VAE with the Simpsons' faces

This task is based on Variation Autoencoder, which is a similar method to Autoencoder with a special variance in the intermediate process.

Reinforcement Learning based on DQN and CNN with the Frozen Lake environment

This task is based on DQN and CNN to deal with the Frozen Lake environment in the gym package. Usually for reinforcement learning, the inputs are state information, but in this project, the inputs are changed into images that show the surrounding around the Agent. It will take a longer time to train the model but with a better generization.

Reinforcement Learning based on Actor and Critic with the Mountain Car Continuous environment

This task is based on Actor and Critic model to deal with Mountain Car Continuous environment in the gym package. Eventually, this model performs not so well.

Reinforment Learning based on DDPG with the Mountain Car Continuous environment

This task is based on DDPG model to solve the task above.

Graph Neural Network based on GCN with the Amazon Product Dataset

This task is based on GCN to implement the node prediction for Amazon Product Dataset.