Search

Specialized function gradient computing hardware could greatly improve the performance of state-of-the-art optimization algorithms, e.g., based on gradient descent or conjugate gradient methods that are at the core of control, machine learning, and operations research applications. Prior work on such hardware, performed in the context of the Ising Machines and related concepts, is limited to quadratic polynomials and not scalable to commonly used higher-order functions. Here, we propose a novel approach for massively parallel gradient calculations of high-degree polynomials, which is conducive to efficient mixed-signal in-memory computing circuit implementations and whose area complexity scales linearly with the number of variables and terms in the function and, most importantly, independent of its degree. Two flavors of such an approach are proposed. The first is limited to binary-variable polynomials typical in combinatorial optimization problems, while the second type is broader at the cost of a more complex periphery. To validate the former approach, we experimentally demonstrated solving a small-scale third-order Boolean satisfiability problem based on integrated metal-oxide memristor crossbar circuits, one of the most prospective in-memory computing device technologies, with a competitive heuristics algorithm. Simulation results for larger-scale, more practical problems show orders of magnitude improvements in the area, and related advantages in speed and energy efficiency compared to the state-of-the-art. We discuss how our work could enable even higher-performance systems after co-designing algorithms to exploit massively parallel gradient computation., Comment: 36 pages, 16 figures

The superconducting charge-phase `Quantronium' qubit is considered in order to develop a model for the measurement process used in the experiment of Vion et. al. [Science 296 886 (2002)]. For this model we propose a method for including the bias current in the read-out process in a fundamentally irreversible way, which to first order, is approximated by the Josephson junction tilted-washboard potential phenomenology. The decohering bias current is introduced in the form of a Lindblad operator and the Wigner function for the current biased read-out Josephson junction is derived and analyzed. During the read-out current pulse used in the Quantronium experiment we find that the coherence of the qubit initially prepared in a symmetric superposition state is lost at a time of 0.2 nanoseconds after the bias current pulse has been applied. A timescale which is much shorter than the experimental readout time. Additionally we look at the effect of Johnson-Nyquist noise with zero mean from the current source during the qubit manipulation and show that the decoherence due to the irreversible bias current description is an order of magnitude smaller than that found through adding noise to the reversible tilted washboard potential model. Our irreversible bias current model is also applicable to the persistent current based qubits where the state is measured according to its flux via a small inductance direct current superconducting quantum interference device (DC-SQUID)., Comment: 42 pages and 7 figures

We show how the measurement induced model of quantum computation proposed by Raussendorf and Briegel [Phys. Rev. Letts. 86, 5188 (2001)] can be adapted to a nonlinear optical interaction. This optical implementation requires a Kerr nonlinearity, a single photon source, a single photon detector and fast feed forward. Although nondeterministic optical quantum information proposals such as that suggested by KLM [Nature 409, 46 (2001)] do not require a Kerr nonlinearity they do require complex reconfigurable optical networks. The proposal in this paper has the benefit of a single static optical layout with fixed device parameters, where the algorithm is defined by the final measurement procedure., Comment: 14 pages, 4 figures, 4 tables