Stochastic computing around kT

In this talk we will review the basic concepts involved in the stochastic modeling of non-linear electronic circuits such as those employed in realistic computing architectures (mainly CMOS circuits). While this class of circuits is usually operated in a deterministic regime, we will discuss how last generation nanoscale fabrication processes enable new design possibilities, where thermal fluctuations are exploited as a resource. In particular, we will discuss the relation between reliability and entropy production in non-equilibrium electronic memories and the design of a binary stochastic neuron. Based on this neuron design, we will discuss the physical implementation of stochastic annealing schemes and artificial neural networks operating at low power. Analogies and possible connections with quantum technologies and computing schemes will be established.