Research [1, 2] shows you can combine XGBoost and CNNs to accurately solve non-image classification problems. In this project, I adapt this model to show a shockingly powerful method for image classification.

This project is my introduction to safe Reinforcement Learning. For this project, I develop a grid-world simulation and an RL algorithm from scratch to experiment with safety RL methods.

Recent research [1, 2, 3] shows how Logic can complement Reinforcement Learning algorithms to make the learning process faster and safer. In this project, I contextualize and critically evaluate the literature on this topic.

I develop two LSTM models based on a dataset of Bitcoin prices. The first model predicts Bitcoin's day-to-day prices. The second model predicts Bitcoin's day-to-day price difference.