Nondeterministic Operation Research Articles

Dual-mode neuron design with deterministic and non-deterministic operations using adiabatic superconductor devices

Abstract In this research, we unveil an innovative strategy in neuromorphic computing by developing a neuron model tailored for the energy-efficient Adiabatic Quantum-Flux-Parametron (AQFP) logic. This model is particularly aimed at enhancing neural network accelerators. Our design of the AQFP-based neuron operates effectively in both deterministic and non-deterministic modes. In deterministic mode, the design relies on superconducting inductive coupling to activate neurons by comparing the sum of AQFP signal currents against a tunable threshold. For non-deterministic operation, we demonstrate how altering specific circuit parameters can correlate these aggregated currents with the non-deterministic operational range of an AQFP current comparator. We verified its versatility and functionality by fabricating varied circuits and conducting extensive tests, confirming its practical application potential. Our work not only showcases the practical implementation of AQFP in neuromorphic computing but also sets a foundation for future advancements in energy-efficient AI hardware.

Modeling Method to Abstract Collective Behavior of Smart IoT Systems in CPS.

This paper presents a new modeling method to abstract the collective behavior of Smart IoT Systems in CPS, based on process algebra and a lattice structure. In general, process algebra is known to be one of the best formal methods to model IoTs, since each IoT can be represented as a process; a lattice can also be considered one of the best mathematical structures to abstract the collective behavior of IoTs since it has the hierarchical structure to represent multi-dimensional aspects of the interactions of IoTs. The dual approach using two mathematical structures is very challenging since the process algebra have to provide an expressive power to describe the smart behavior of IoTs, and the lattice has to provide an operational capability to handle the state-explosion problem generated from the interactions of IoTs. For these purposes, this paper presents a process algebra, called dTP-Calculus, which represents the smart behavior of IoTs with non-deterministic choice operation based on probability, and a lattice, called n:2-Lattice, which has special join and meet operations to handle the state explosion problem. The main advantage of the method is that the lattice can represent all the possible behavior of the IoT systems, and the patterns of behavior can be elaborated by finding the traces of the behavior in the lattice. Another main advantage is that the new notion of equivalences can be defined within n:2-Lattice, which can be used to solve the classical problem of exponential and non-deterministic complexity in the equivalences of Norm Chomsky and Robin Milner by abstracting them into polynomial and static complexity in the lattice. In order to prove the concept of the method, two tools are developed based on the ADOxx Meta-Modeling Platform: SAVE for the dTP-Calculus and PRISM for the n:2-Lattice. The method and tools can be considered one of the most challenging research topics in the area of modeling to represent the collective behavior of Smart IoT Systems.

Quantum circuits with many photons on a programmable nanophotonic chip.

Growing interest in quantum computing for practical applications has led to a surge in the availability of programmable machines for executing quantum algorithms1,2. Present-day photonic quantum computers3-7 have been limited either to non-deterministic operation, low photon numbers and rates, or fixed random gate sequences. Here we introduce a full-stack hardware-software system for executing many-photon quantum circuit operations using integrated nanophotonics: a programmable chip, operating at room temperature and interfaced with a fully automated control system. The system enables remote users to execute quantum algorithms thatrequire up to eight modes of strongly squeezed vacuum initialized as two-mode squeezed states in single temporal modes, a fully general and programmable four-mode interferometer, and photon number-resolving readout on all outputs. Detection of multi-photon events with photon numbers and rates exceeding any previous programmablequantum optical demonstration is made possible by strong squeezing and high sampling rates. We verify the non-classicality of the device output, and use the platform to carry out proof-of-principle demonstrations of three quantum algorithms: Gaussian boson sampling, molecular vibronic spectra and graph similarity8. These demonstrations validate the platform as a launchpad for scaling photonic technologies for quantum information processing.

Design, Test, and Repair of MLUT (Memristor Look-Up Table) Based Asynchronous Nanowire Reconfigurable Crossbar Architecture

The recently proposed nanoscale asynchronous crossbar architecture based on memristor-based look up table (MLUT) combines the advantages of nanoscale memristor crossbar technology and clockless logic paradigm for viable nanoscale computing. Potential technical merits of the proposed MLUT architecture includes: 1) better manufacturability due to structural simplicity and regularity; 2) improved robustness over PVT (process-voltage-temperature) variations; 3) event-driven low-power/noise asynchronous operation; and 4) encoding-level logic inversion. In spite of having numerous merits over the clocked counterparts and previous asynchronous designs, it is bound to have inevitable defects and faults due to nondeterministic and unconventional nanoscale assembly and operation. In order to overcome defect issues in the proposed MLUT-based nanoscale asynchronous crossbar architecture, there is a need to develop efficient design, test, and repair techniques. Typical approach so far has been to test every crosspoint on each crossbar MLUT exhaustively; this is not only laborious but is also prohibitively time and space consuming for designs involving large number of MLUTs. This paper introduces a novel testing scheme based on “Divide and Conquer” approach to efficiently locate the defective memristors in a MLUT. The proposed testing scheme leverages upon a special current additive property of the memristor-based multiplexer. It performs binary isolation of regions, reducing the search space by half whenever applicable. Numerical simulations clearly demonstrate that the approach is generic, deterministic, and scalable. A faster MLUT programming technique and a repair technique utilizing partially defective MLUTs are also proposed and extensively validated through parametric simulations.

A representable approach to finite nondeterminism

We reformulate denotational semantics for nondeterminism, taking a nondeterministic operation V on programs, and sequential composition, as primitive. This gives rise to binary trees. We analyse semantics for both type and program constructors such as products and exponential types, conditionals and recursion, in this setting. In doing so, we define new category-theoretic structures, in particular premonoidal categories. We also account for equivalences of programs such as those induced by associativity, symmetry and idempotence of V, and we study finite approximation by enrichment over the category of ω-cpos with least element. We also show how to recover the classical powerdomains, especially the convex powerdomain, as three instances of a general, computationally natural, construction.

Evaluation of operators defined by regular schemas

We consider different techniques of organizing the evaluation of operators defined by regular schemas. The proposed techniques are based on an abstract machine (a discrete process generator) with nondeterministic operation.