Deep Reinforcement Learning with Python

Topic 1 Introduction to Reinforcement Learning

• Fundamental Concepts of Reinforcement Learning (RL)
• Types of RL Algorithms
• Applications of RL
• Markov Decision Process

Topic 2 OpenAI Gym and Stable Baselines

• Introduction to OpenAI Gym
• Install OpenAI Gym and Stable Baselines
• Create Agent and Policy on Gym

Topic 3 Value Based Q-Learning

• Overview of Value Based Learning
• Value Functions and Bellman's Equations
• Exploration Strategies
• Q-Learning Algorithm
• SARSA Algorithm
• Deep Q Network (DQN) Algorithm

Topic 4 Policy Based Learning

• Overview of Policy Based Learning
• Policy Network
• Policy Gradient Algorithm

Topic 5 Advanced RL Algorithms

• Actor-Critic A2C/A3C Algorithms
• Proximal Policy Gradient (PPO/PPO2)

Topic 6 Advanced Stable Baselines Techniques

• Custom Policy Networks
• Callbacks and Tensorboard
• Reinforcement Learning on Atari Games
• Reinforcement Learning on MuJoCo

Topic 7 Brief Introduction to Model-Based Learning

• What is Model-Based Learnings
• Monte Carlo Tree Search
• Brief Overview of AlphaZero

Prerequisite

This is an intermediate course. The following knowledge is assumed:

• Basic Python

Software Requirement

Please install the following software prior to the class

1. Pycharm : - Install Pycharm (https://www.jetbrains.com/pycharm/download/)

2 . Install Tensorflow on Mac

Please follow this guide to install Tensorflow on Mac https://www.tensorflow.org/install/install_mac

Alternatively, you can enter the following commands on your Mac terminal

pip3 install tensorflow

3 . Install Tensorflow on Window

Please follow this guide to install Tensorflow on Window https://www.tensorflow.org/install/install_windows

HRDF Funding

Please refer to this video https://youtu.be/Kzpd-V1F9Xs

1-     HRD Corp Grant Helper

How to submit grant applications for HRD Corp Claimable Courses

2-     Employers are required to apply for the grant at least one week before training commences.

Employers must submit their applications with supporting documents, including invoices/quotations, trainer profiles, training schedule and course content.

3-     First, Login to Employer’s e-TRIS account -https://etris.hrdcorp.gov.my

Second, Click Application

4-     Click Grant on the left side under Applications

5-     Click Apply Grant on the left side under Applications

6-     Click Apply

7-     Choose a Scheme Code and select HRD Corp Claimable Courses: Skim Bantuan Latihan Khas. Then, click Apply

8-     Scheme Code represents all types of training that suit the requirements provided by HRD Corp. Below are the list of schemes offered by HRD Corp:

9-     Select your Immediate Officer and click Next

10-  Select a Training Provider, then click Next

11-  Please select a training programme from the list, then key in all the required details and click Next

Select your desired training programme.

Give an explanation on why the participant is required to attend the training. E.g., related to their tasks/ career development, etc.

Explain the background and objective of this training.

Select a relevant focus area. For Employer-Specific Courses, select ‘Not Applicable’.

12-  If the training programme is a micro-credential programme, you are required to complete these 3 fields. Save and click Next

Insert MiCAS Application number

13-  Based on the nine (9) pillars listed below, HRD Corp Focus Area Courses are closely tied to support government initiatives towards nation building. As such, courses offered through the HRD Corp Focus Areas are designed to provide the workforce with skills required for current and future demands. Details of the focus areas are as follows:

14-  Please select a Course Title and Type of Training

15-  Select the correct type of training according to the actual type of training, or as mentioned in the training brochure:

16-  Please key in the Training Location and click Next

17-  Please select the Level of Certification and click Next

18-  Please follow the instructions and key in trainee details

19-  Click Add Batch, then click Save

20-  Click Add Trainee Details

21-  Please key in all the required details, then click Add

22-  Click Add if there are more participants. Once done, click Save

23-  Click Next

24-  Please key in the course fees and allowance details, then click Save

25-  Estimated cost includes the course fees/external trainer fees, allowances, and consumable training materials. Please comply with the HRD Corp Allowable Cost Matrix.

26-  Select Upfront Payment to Training Provider and key in the percentage from 0% to 30%. Then, click Save and Next

27-  Complete the declaration form and select a desired officer

28-  Add all the required documents, then click Add Attachment. Then, click Save and Submit Application

29-  Once the New Grant Application is successfully submitted, the Grant Officer will evaluate the application accordingly. The application may be queried if additional information is required.

The application status will be updated via the employer’s dashboard, email, and the e-TRiS inbox.

• Machine Learning Engineer
• Robotics Engineer
• Game Developer (AI-focused)
• AI Research Scientist
• Data Scientist (branching into RL)
• Autonomous Systems Developer
• Simulation Engineer (using RL)
• Optimization Specialist
• AI Product Manager (oversight on RL projects)
• Control Systems Engineer (using RL)
• Finance Quant (using RL for trading strategies)
• NLP Engineer (using RL for certain applications)
• Recommendation System Developer (using RL)
• AI Solutions Architect
• Drone Algorithm Developer.

Nero: Nero is dean of Penang School of AI, which is AI learning and sharing community in Northern Malaysia. He is Msc in Computer Imaging in Universiti Sains Malaysia. His research is mainly focusing in Video Analytic and Multiple object tracking. At the same time, he is developing computer vision application in WyseTime, a tech-startup company focus on video analytic application using deep learning technology.

Lee Cheong Loong: Lee Cheong Loong, Manager with 23 years working experience in multiple role and department, He completed HRD Corp Train the Trainer programme, HRD Corp Accredited Trainer, Microsoft Certified Trainer and CPFA Citizen Data scientist Trainer programme. with Professional certificate in Big Data & Analytics, Microsoft Office Specialist -Excel 2016 and Tableau Desktop Specialist. He also deliver training for R & Python programming, Excel Dashboard for Business analysis, Data Visualization with Tableau, and Microsoft PowerBI, and Citizen Data Scientist (OpenCertHub).

Dr Azam: Dr Azam is an aerospace engineer with a passion on machine learning. He self-taught himself R statistical methods and has developed a machine learning course specifically for mechanical engineers; curating example case studies where traditional engineering analysis benefitted from machine learning techniques. Currently, he is involved in analyzing data from an oil platform as part of a large predictive maintenance package. He is also interested in mobile apps development and has consulted a few businesses to modernize their point of sale system using tailor-made android apps. He has supervised final year projects related to android mobile apps and one of the projects has won silver award for the best project overall.