QuaterMaster

Overview

For the approach to the machine learning, we adopted a way to use physical theory like quantum annealing and neural network to enforce existing device and /or tool solutions such as conventional computing way, GPU and FPGA.

Problems to solve

Machine Learning
Deep Learning
Crytography
Data Analysis
Finite Element Method

Problems to research

Automatic layout and wiring system for integrated circuit
Route search method based on quantum annealing
Finding optimal formulation for drug and chemical substance development

About Quantum Annealing Simulation

Overview

We are simulating quantum annealing on a classical computor based on physical theory to discretization the Schrodinger equation.

Path Integral

We discretize the Schrodinger equation using path integral to several Trotter Slice and calcurating the interaction between adjoining slices.

Annealing

We control two way of annealing (Temperature and Quantum effect) by controlling two parameter of beta and gamma.

About Machine Learning adopted to logical circuit

Overview

In a basic physical theory, many phenomenon is written in Ising model . Basically these Ising Model has {-1,1} value of each spins.
To interpret these Ising model to Quantum annealing logical circuit, we are using QUBO to transform {-1,1} to {0.1}.

Bit notation

To write a decimal number, we have to use some binary qubit to write these numbers.

Logical Operations

For expample, to solve prime factorization ,we need to solove (N-pq)^2.Each of "p" and "q" is decimal number by binary qubit.
N^2 + 2pqN + (pq)^2 has calcuration of interaction over 2 factors. To solve 3 or 4 interaction of qubits, we need to decomposite these to 2 interaction.

QuaterMaster{logical simulator based on Quantum Annealing}

Machine Spec
Tuning Mac Pro for Quantum annealing
-3.7 GHz quad Core Intel Xeon E5
-12Gb 1,866Mz DDR3 ECC memory
-Dual AMD FirePro D300 (2GB GDDR5 VRAM)
-256GB flash storage (PCIe)

About

It has a sepcific logical circuit simulator based on quantum annealing algorithm.

Purpose

To evaluate a problem which can solve on a quantum annealing adiabatic computer.

Problems to solve Algorithm
SAT Quantum Annealing
Prime factorization Quantum Monte Carlo
Clustering Metropolis
etc... Embedded Graph