Exciting example of a quadruped robot using vision and a convolutional neural network to continuously replan its steps while facing external challenges such as slippery surfaces and mistakes in foot placement. 

Learn more about NEURAL NETWORKS  here

 MACHINE LEARNING 

Some useful resources for students of this field.

A computer is said to learn:

from experience E 

with respect to task T 

with some performance measurement P

 if its performance on T, as measured by P, improves with experience E.

-Tom Mitchell -

Here’s a well demonstrated overview of a neural network. Programming this example is a delightful exercise and a “must do” for any student of neural networks.

For another approach to designing a network model for continuous processes, see the Articles tab on this site for the article "Neural Ordinary Differential Equations". The authors introduce a family of deep neural network models (ODE nets) which do not use a discrete sequence of hidden neural net layers with nodes and connections.

 An overview of a Convolutional Neural Network (CNN/ConvNet)

 Gradient Descent optimization demonstration - a bit easier than conceptualizing with contour...