Boltzmann-Machine

Boltzmann machine to solve the data association problem. BM is a symetrically interconnected network of neuron-like units. Those units make stochastic decisions whether to be on or off . This means that the Boltzmann machine accepts or rejects the elemental hypotheses represented by their units at each iteration. The aim of the Boltzmann machine is to find probable configuration of units by minimizing an energy function. Thus, it will have a stationary distribution which is equal to the distribution over valid joint association events. To this end, the Boltzmann machine uses the simulated annealing schedule. The Boltzmann machines are stacked to form a parallel computational organization. This leads to the computationally efficient and accurate solutions to constraint satisfaction tasks.