An intuitive Intro to Boltzmann Machines
This model doesn’t have an output layer. Everything is connected to everything. There’s no direction in these connections.
Boltzmann machines are fundamentally different to all other algorithms.
They don’t expect the data but rather generate the data. For Boltzmann machine every node is same. We want to use our training data to feed the Boltzmann Machines, as the inputs to help it adjust the weights of the system accordingly so that it actually resembles our system.
It’s a stochastic or generative deep learning model.
Once the learning is done, the Boltzmann machine understands how all these parameters interact with each other, or what constraints should exist between them in order for the system that we are modeling. An example would be the Boltzmann Machine that monitors the Nuclearn Powerplant. Learning through good examples, it understands how the system works in it’s natural state.
