Description
In this course, you will learn :
- Fundamental concepts and terminologies used in deep learning.
- Understand the significance of deep learning techniques.
- Simple models such as perceptrons before moving on to more complex but powerful deep learning models.
- How to code simple and complex deep learning models in NumPy, a powerful Python library, and Keras, a cutting-edge Python deep learning library.
Syllabus :
1. Introduction to Deep Learning
- Fundamentals of Machine Learning
- Machine Learning Paradigms
- What is Deep Learning?
- Programming Paradigm
2. Simple Perceptron Models in NumPY
- Introduction to Perceptron
- Coding the Perceptron Forward Propagation
- Discrete vs. Continuous Prediction
- The Error Function
- Challenge: Scaling Error Up to Multiple Data Points
- Solution Review: Scaling Error Upto Multiple Data Points
- Optimizing the Perceptron Output
- Gradient Descent: Stochastic vs. Batch Update
- Gradient Descent: The Stochastic Update
- Gradient Descent: The Batch Update
- Perceptrons as Logical Operators
- Problems with Gradient Descent and the Fix
3. Towards Deep Neural Networks in NumPY
- Introduction to Non-Linear Boundaries
- Neural Network
- Feed Forward Propagation
- Back Propagation
- Activation Functions
- Deep Neural Network
4. Building Deep Learning Models with Keras
- Introduction to Keras
- Keras Workflow
- Read and Explore the Data
- Specify Model Architecture
- Model Compilation
- Fitting a Model
- Model Evaluation
- Using Models
- Regression Models
5. Fine-tune Keras Model
- Introduction to Model Optimization
- Model Validation
- Model Capacity