Boltzmann Machines are bidirectionally connected networks of stochastic processing units, i.e. units that carry out randomly determined processes.
Read the full article for complete details and intuition behind Restricted Boltzmann Machines and join the discussion now
A Boltzmann Machine can be used to learn important aspects of an unknown probability distribution based on samples from the distribution. Generally, this learning problem is quite difficult and time consuming. However, the learning problem can be simplified by introducing restrictions on a Boltzmann Machine, hence why, it is called a Restricted Boltzmann Machine .
Restricted Boltzmann Machines has found applications in Quantum Computing and is commonly used in dimensionality reduction, classification, feature learning, topic modelling , recommendation systems and music generation.
This is a companion discussion topic for the original entry at http://iq.opengenus.org/restricted-boltzmann-machine/