Description
In this course, you will learn :
- To solve physical problems in simulated environments, configure and use the Unity Machine Learning Agents toolkit.
- About neural networks, supervised and deep reinforcement learning (PPO).
- In Unity, use ML control techniques to teach a go-kart to drive around a track.
Syllabus :
1. Self-driving cars
-
Self-driving cars
2. Unity Machine Learning Agents
- Self-driving go-kart project in Unity
- Machine Learning Brains
- Control Scripts
- Setup the ML-Agents Toolkit
3. Traditional Control
- Traditional Control
- PID Controller
- Tuning
- Heuristic brain in Unity
- Cross-track error
- Testing the PID
- Improvements
- Model-based control
- Onto Machine Learning
4. Imitation Learning
- Why Machine Learning
- What kinds of learning
- Neural Networks
- NN Details
- Training a NN
- Optimizer
- Convolutional layers
- Transfer learning
- Imitation learning in Unity
- Training the go-kart via IL
- Testing the drive
- Tips on imitation learning
5. Reinforcement Learning
- Reinforcement Learning
- Nomenclature
- Initial state
- Training a policy
- The PPO algorithm
- Evolutional Strategies
- Reward
- Training the go-kart with RL
- Tensorboard analysis
- Testing results