Based on our discussions and further thinking, I have outlined what needs to be done right now. First, Chapter one will give a brief overview of deep learning and environment setup. Then, in order to gradually establish students’ understanding of deep learning, Chapter 2 and Chapter 3 will explain the basic concepts step by step through examples. Especially, In Chapter 3, we can do a physical experiment, involving every student, to simulate the propagation process of the neural network. Finally, in chapter 4, students will learn how to write a two-layers neural network to complete a specific task without applying the TensorFlow library. Because it’s an extremely compact and lightweight model, we can run it directly on the Raspberry Pi which could be a lab too.
After that, I will further demonstrate more complex deep learning projects. For example, as we mentioned before, investigating News data adopting TensorFlow to discovery the causal relationship.