Improved representations and hardware implementation of the Helmholtz Machine

Probabilistic computing forms a relatively new computational style, of significant practical interest because stochastic behaviour is common and must be taken into accountin in biological and other real-world processes. We examine a particular stochastic ANN architecture, the Helmholtz Machine, investigating its characteristics, with particular respect to its wake-sleep training algorithm, and showing how its representational power might be increased. We also explain how we are implementing a simple version of the machine in analogue VLSI hardware.