What is a Boltzmann machine?

A stochastic recurrent network of symmetrically connected binary units whose on-probability depends on net input and temperature

A deterministic feed-forward network with no hidden units

A supervised single-neuron classifier

A clustering network using winner-takes-all

Why is simulated annealing used?

To avoid getting trapped in local energy minima

To speed up forward passes only

2070 Magh · 1 mark

Define Boltzmann machine

Short answer

A Boltzmann machine is a stochastic recurrent network of binary units with symmetric weights. Units switch on probabilistically based on energy and temperature. Simulated annealing helps avoid local minima.

Define Boltzmann Machine.

A Boltzmann machine is a stochastic recurrent neural network of symmetrically connected binary units (visible and hidden). Each unit switches on with a probability that depends on its net input and a temperature T, and the network is trained so that its energy function reaches a minimum for the desired patterns; simulated annealing is used to avoid local minima.

✓ Answer checked by Claude Opus 5.5 (high effort)

Machine learning; Boltzmann machines · Artificial Intelligence, chapter 6

All Artificial Intelligence chapter 6 questions →

Define Boltzmann machine

Related questions