Deep Learning @ CYCU
Course Content

Deep learning is one of the mainstreams in artificial intelligence, as well as, in data science. We usually leverage deep learning techniques in the complicated problems or data with unknown characteristics. Hence, the major advantages of deep learning framework are designed for automatically identifying possible features to minimize the outcome of loss function. In this semester, we will elaborate the core concept and programming of major deep learning algorithms. All students have to conduct the deep learning algorithms to demonstrating their understanding in solving the real-world problems.

Course Intro.

01 :: Course Introduction

Contents: (1) Course intro (2) Why do you need to take this course? (3) What will you learn from this course? (4) Syllabus (5) Grading policy

Slide

Machine Learning & Deep Learning

02 :: Machine Learning & Deep Learning

Contents: (1) Artificial intelligence (2) Machine learning (3) Deep learning (4) Machine learning & deep learning (5) Optimization (7) Potential issues (8) Paper reading

Slide

Neuron Network Basis

03 :: Neuron Network Basis

Content: (1) Model formulation (2) Single nueron (3) Hidden layer (4) Activation function (5) Linearity and non-linearity (6) Model parameters (7) Loss function

Slide

Gradient Descent

04 :: Gradient Descent

Content: (1) Finding the best parameter set (2) Loss optimization (3) Global minima and local minima (4) Gradient descent (5) Stochastic gradient descent (6) Mini-batch stochastic gradient descent (7) Initial condition (8) Learning rate

Slide

Backpropagation

05 :: Backpropagation

Content: (1) Forward and Back Propagation (2) Chain Rule (3) 𝝏𝑪(𝜽)/𝝏𝒘 (4) 𝝏𝑪(𝜽)/𝝏𝒛 (5) 𝝏𝑪(𝜽)/𝝏𝒛𝒍 = 𝜹𝒍 (6) Back propagation (7) Summary

Slide

Keras - RNN & LSTM

06 :: Keras - RNN & LSTM

Content: (1) Recurrent neural network (2) Long-short term memory

Code