Click at any point on the panel to plot the gradient vector. Can you find the local maxima and minima of the function?
Principal Components Analysis can be used to reduce the dimension of points by projecting them onto a smaller dimensional space. Here are some examples of 3D shapes being projected onto their first 2 principal components.
Neural Networks key processes involve minimizing a loss function using gradient descent and updating the parameters using backward propagation. This interactive applet aims to classify points in two categories (close to the origin and far from the origin). Because of the layout, without a prior transformation this dataset is not suited for other classification techniques like logistic regression. In it, you can take steps of the gradient descent, and see how the weights and biases of the neural network, its classification and the loss function vary in each step.