The brain and the computer

Summary form only given. Exactly fifty years ago, when he published "The computer and the brain" in 1957, John von Neumann foresaw computer designers benefiting from basing their designs on the brain. Interest in this topic has been renewed by the convergence in properties of transistors and ion-channels (the brain's transistors). At ten nanometers, a transistor's channel becomes so narrow that an electron trapped by a dangling bond at the surface (an unavoidable atomistic defect) blocks electron flow. The current turns off and on at random, as trapping and detrapping occurs stochastically. Such stochastic behavior, which corrupts the deterministic on/off states computers rely on to perform binary arithmetic, is also displayed by ion-channels. At under a nanometer in size, an ion-channel's gate is agitated by thermal forces, opening and closing randomly. I will describe efforts in the neuromorphic engineering community to explore how the brain computes with stochastic devices by emulating an ion-channel's ionic current directly with a transistor's electronic current. While a present-day transistor, at a hundred-nanometers wide, corresponds to a small population of ion-channels, not a single ion-channel, this analog approach provides an extremely efficient method to simulate the brain while at the same time laying the groundwork for building brain-like computers out of next decade's nanotransis-tors. John von Neumann was prescient in anticipating that computer designers could profit by modeling features of the brain in their designs - even though he did not foresee the remarkable device-level convergence.