Research

Demos, Research and Practice for Thesis

---

Demos

1. DeepQ-CartPole

Notebook with OpenAIGym cartpole example. Simple DQN implementation with notes and videos.

2. Neural Nets

Simple Neural Network: Notebook using Keras and MNIST dataset. Trained model and outputs summary and visualization.

3. DeepRL Course Berkeley

Independent study using Berkeley Deep RL course materials and assignments. Assignment 1: Imitation Learning Assignment 2: Policy Gradients

4. DeepRL-OpenAIGym

Deep Reinforcement learning (actor-critic) with prioritized experience replay. Paper Implementation Continuous Control with deep Reinforcement Learning, using Code and Tutorial From Paperspace blog. Updated to include priority experience replay. Other sources include: Array backed Sum Tree code, Priority experience replay, DDDQN with prioritized experience replay, Memory.

4. Open AI Gym

Simulation toolkit for comparing reinforcement learning algorithms.
GYM Github

---

Research

---

Practices

1. PyTorch: tutorial

---

Links

Informational

1. Deep Q Learning
2. Soft Actor Critic Deep RL
3. Soft Learning Code Implementation
4. Parameter Noise
5. Berkeley DeepRL Course Github Code

Tools

1. Conda virtual environments
2. Guide to Virtual Environments
3. Install Mujoco for Windows
4. Baxter Grasping DL Implementation