Computing using unknown values

Patent metadata

Field	Value
Publication number	US-10360285-B2
Application number	US-201715802342-A
Country	US
Kind code	B2
Filing date	Nov 2, 2017
Priority date	Nov 2, 2017
Publication date	Jul 23, 2019
Grant date	Jul 23, 2019

Abstract

A method of computing includes defining a first atomic random variable (ARV) and first random variable (RV) in a programming language system. The first ARV having a non-deterministic value of either zero according to a second probability or one according to a first probability. A sum of the first probability and the second probability is one. A covariance of the first ARV and a second ARV is zero. The first RV has a first indefinite value at a first definite probability and includes a polynomial of one or more atomic random variables (ARVS) that includes the first ARV. The method includes executing a computer instruction that includes a mathematical operation involving the first RV as a basic data type and produces a second RV having a second indefinite value at a second definite probability, represents a result distribution, and tracks a response to the one or more ARVS.

First claim

What is claimed is:

1. A method of quantum-inspired computing, the method comprising:
defining a first atomic random variable (ARV) in a programming language system, the first ARV having a non-deterministic value of either zero or one, a first probability of having a value of one, and a second probability of having a value of zero; a sum of the first probability and the second probability is one; and a covariance of the first ARV and a second ARV is zero;
defining a first random variable (RV) in the programming language system, the first RV having a first indefinite value at a first definite probability and includes a polynomial of one or more atomic random variables (ARVS) that includes the first ARV; and
executing a computer instruction that includes a mathematical operation involving the first RV as a basic data type, the executing producing a second RV that has a second indefinite value at a second definite probability, represents a result distribution, and tracks a response to the one or more ARVS.

2. The method of claim 1 , wherein:
a mean of the first ARV is equal to the first probability;
a mean square of the first ARV is equal to the first probability;
an n-th moment of the first ARV is equal to the first probability for all values of n;
a variance of the first ARV is equal to a product of the first probability and the second probability; and
an m-th power of the first ARV is equal to the first ARV for all values of m.

3. The method of claim 1 , wherein the first ARV is defined according to ARV expressions:
X i ∈{0,1};

0 ≤p i ≤1;

Pr[ X i =1]= p i ;
Pr[ X i =0]=1 −p i ; and
Cov( X i ,X k )=0 for i≠k ; in which:
i and k represent indexing variables;
p i represents a first probability indexed according to the indexing variable i;
X i represents an ARV indexed according to the indexing variable i;
X k represents an ARV indexed according to the indexing variable k;
Pr[ ] represents a probability function; and
Cov( ) represents a covariance function.

4. The method of claim 3 , wherein entropy of the ARV is approximated according to an approximate entropy expression:
S [ X i ]≃ log 2−2 q i 2 ; in which:

S[X i ] represents the entropy,
q i represents p i −½, and
log 2 represents a natural log of 2 base Napier's constant (e).

5. The method of claim 3 , wherein:
the first random variable is a quadratic random variable (QRV); and
the QRV is defined according to the QRV expressions:

Y
A

≡

X
→

T

⁢
A
⁢

X
→

;

X
→

≡

[

X
0

⁢

⁢

X
1

⁢

⁢

X
2

⁢

⁢
⋯
⁢

⁢

X
m

]

T

;

A
≡

[

a

0
,
0

a

0
,
1

a

0
,
2

…

a

0
,
m

a

1
,
0

a

1
,
1

a

1
,
2

…

a

1
,
m

a

2
,
0

a

2
,
1

a

2
,

Assignees

• Fujitsu Ltd

Inventors

• Hidaka Yasuo

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06F17/13

  Differential equations (using digital differential analysers G06F7/64) · CPC title

• G06F7/70

  using stochastic pulse trains, i.e. randomly occurring pulses the average pulse rates of which represent numbers {(conversion of analogue signals into stochastic pulse trains and vice versa H03M1/04)} · CPC title

• G06F7/544

  for evaluating functions by calculation {(G06F7/4824 takes precedence)} · CPC title

• G06F7/58

  Random or pseudo-random number generators · CPC title