amiaty / Boltzmann-Machine-Matlab

Simple implementation of the standard Boltzmann Machine model with MATLAB

Boltzmann Machine (BM)

• The code is an implementation within the MATLAB of the Standard Boltzmann Machine model.
• The Implementation is based on Ackley, David H., Geoffrey E. Hinton, and Terrence J. Sejnowski 's paper A Learning Algorithm for Boltzmann Machines, Cognitive Science 9.1 (1985).

Summery

Boltzmann Machines are very similar to Hopfield Networks (HNs) where some units are marked as input and others consider as hidden units. Input units become output as soon as each hidden unit update its state (during training, a Boltzmann Machine updates its cells one by one (not in parallel) in a random order (same as Hopfield Networks)).

Getting Started

Dependencies

• MATLAB (Almost any versions)

Installing

• Download repository and extract it.
• Open folder in MATLAB. (No needs for additional files or data)

Executing program

• Run test.m to see the results of the paper's experiments.

Note

The full model and code logic existed in the single class file BoltzmannMachine.m

Authors

• Kamaledin Ghiasi-Shirazi
• Amir Ahooye Atashin

License

This project is licensed under the MIT License - see the LICENSE file for details