Oliver Thewalt

    Oliver Thewalt

    Theoretical Physics | Quantum Biology | Dark Matter Research Cluster

    Massively parallel neural circuit use complete wafer

    by Georg Drakenberg, Athens

    Fast Analog Computing with Emergent Transient States and BrainScaleS Neuromorphic Systems Board Prototype. Brain-inspired Multiscale Computation in Neuromorphic Hybrid Systems. In the first layer (down), massively parallel neural circuit use complete wafer to exploit inherent fault tolerance and asynchronous pulse transmission. In the secont layer (up, covered by an aluminum housing), a packet based digital communication protocol for long-range spike transmission, network setup, monitoring and control. That's a Neural VLSI-based Neuromorphic Platform of Reverse Engineering of the Brain, BrainScaleS Project of Reverse Engineering of the Brain, University of Heidelberg (directed by Dr. Karlheinz Maier et al., started on 1 January 2011). The roadmap of international Neuromorphic Engineering Projects includes, architectures that based on Cellular Arrays with Defect Tolerance adn Biologically Inspired. The Logic Code System that will be used includes Rapid Single Flux Quantum Technology (RSFQ), 1D nanostructures (quantum dots), Resonant Tunneling, Single Electron Transistors and Spintronics. For Memory, Phase Change Floating Point DRAM, SDRAM, Nano-Field Gate Arrays, and Molecular Insulator Resistance Changed Components. So, the Ultra Thin Bodies coupled with Single Gate FET and SET, will give a new form of Non-Classical CMOS Neuromorphic Computers (in exaperimental stage for some time).