Dive into the world of deep learning with this comprehensive 7-hour course focused on using Keras and TensorFlow in R. Begin with regression as a foundation, exploring linear regression and its connection to shallow neural networks. Progress through essential concepts like linear algebra, derivatives, and gradient descent. Learn to implement logistic regression and simple neural networks. Gain hands-on experience with R for deep learning, including building deep neural networks using Keras. Explore crucial topics such as bias, variance, regularization, and dropout. Enhance your skills by comparing models with tfruns and investigating sequential models in Keras. Delve into advanced subjects like cross-entropy loss functions and convolutional neural networks, culminating in a practical application for skin lesion analysis using Keras in R.
Overview
Syllabus
Regression as a first step in deep learning.
Linear regression as a simple learner.
Basic linear algebra for deep learning.
Basic derivatives for deep learning.
Gradient descent.
Linear regression as a shallow neural network.
Logistic regression as a network.
Simple neural network.
Introduction to R for deep learning.
Example of a deep neural network using Keras in R.
Bias and variance in deep learning.
Regularization in deep learning.
Dropout in deep learning.
Regularization and dropout using Keras for R.
Improving learning in deep neural networks.
Using tfruns to compare models.
Exploring sequential models in Keras for R.
The cross entropy loss function.
Deep neural networks for regression problems.
Introduction to convolutional neural networks.
Example of a convolutional neural network.
Convolutional neural network using Keras for R - SKIN LESIONS.
Taught by
Dr Juan Klopper
