Neural Networks

Boltzmann Machine

A stochastic network that learns a probability distribution over its inputs.

Units switch on and off probabilistically, and training adjusts the weights so the distribution the network settles into matches the distribution of the data.

The unrestricted version is impractically slow; the restricted variant, with no connections within a layer, was for years the standard way to pre-train deep networks before better initialisation made that step unnecessary.

Also in Neural Networks

Perceptron Backpropagation Feed Forward Hopfield Network Self Organising Maps Deep Belief Network

Apply

Begin the first module

Become AI native, it is the real deal today, and if it is not for you, you have lost nothing but learnt a new skill.

Apply