02. Program Structure
Term 2 Program Structure
Term 2 is divided into three parts. You'll begin with topics in deep learning, and then move into reinforcement learning. Then, the term will culminate in a capstone project of your choosing, that applies the machine learning techniques and algorithms you have learned.
Deep Learning
You've learned what a neural network is in the Supervised Learning section of Term 1. In this first section of Term 2, we'll go even deeper, and study convolutional neural networks, and their applications in image recognition. Other cutting-edge techniques such as transfer learning will be studied, so you'll be at the forefront of the technology!
Reinforcement Learning
Some actions, such as teaching a robot how to walk, or how to play chess, require much more than a supervised or unsupervised learning algorithm. They require the agent to interact with an environment repeatedly and learn from its experiences. This is what reinforcement learning is about. In this section you'll learn how to implement different reinforcement learning algorithms, and apply them to a variety of tasks such as navigating a cab in an environment, or playing atari games.
Then we'll go into the cutting edge algorithms like deep Q-networks or policy gradients, to learn much more complex tasks such as controlling a quadcopter in the air.
Capstone Project
This section has two phases. The first is the Capstone Proposal, during which you will draft a proposal outlining the domain of the problem you would like to solve, and your approach. This is followed by the Capstone Project: here, you will leverage your newly-learned skills to solve the problem—as outlined in your proposal—by applying machine learning algorithms and techniques.