Image Recognition with Neural Networks From Scratch
- 3 hours on-demand video
- 5 downloadable resources
- Full lifetime access
- Access on mobile and TV
- Certificate of Completion
Get your team access to 4,000+ top Udemy courses anytime, anywhere.Try Udemy for Business
- Write a Python program that recognizes images from scratch without using any libraries!
- Understand A Neural Network is.
- Understand some important mathematical prerequisites such as functions and their computational graphs.
- Understand conceptually what a derivative and a gradient is to fully appreciate the Gradient Descent Algorithm.
- Understand the Gradient Descent Algorithm, the central algorithm in machine learning with Neural Networks.
- Understand Backpropagation and its importance in computing gradients.
- Be able to implement the full Python program in 50 lines of code that recognizes images.
Students will learn:
- About the MNIST dataset images of handwritten digits.
- Each image is a 28x28 greyscale. Flatten this 2D array into a 1D vector of dimension 784 and store as a 1D Numpy array.
- The score function that maps each image of pixels to a vector of class scores. The class with the highestscore is the classification of the image. The objective of this class is to find the parameters for this score function.
- The score function is simply a composition of matrix multiplication and addition and the logistic function.
- Learnable parameters of this score function needs to be optimized to find the best score function. This machine learning process is data driven.
- Feed the computer many images and the computers will learn the best parameters that best describe the images.
- Some basic knowledge of Python.(Supplemental "Crash Course" resources are provided to review/learn Python.)
- Some basics knowledge of Numpy.(Supplemental "Crash Course" resources to review/learn Numpy.)
- Some high school precalculus.
This is an introduction to Neural Networks. The course explains the math behind Neural Networks in the context of image recognition. By the end of the course, we will have written a program in Python that recognizes images without using any autograd libraries. The only prerequisite is some high school precalculus. Although the prerequisite is minimal, we will discuss many advanced topics including:
1) functions and their computational graphs.
2) neural networks
3) conceptually understand the derivative and the gradient.
4) gradient descent and backpropagation
5) the multivariable chain rule
6) mini-batch gradient descent
- Beginner Developers who wish to understand Neural Networks.
- Any math enthusiast who wishes to understand how matrix multiplication and the exponential function are the only two functions needed to recognize images!
- Any student who wishes to see one of the most useful and powerful application of high school math!