Optical computing machine
The purpose of the present invention is to provide an optical computer having high scalability and capable of reducing noise. This optical computer is characterized by including: an input layer which accepts an input signal as a computer; a reservoir layer which responds to a signal input from the input layer, and in which using a plurality of degenerate optical parametric oscillator (DOPO) pulses of a DOPO that are generated by an optical phase-sensitive amplifier (PSA) based on pump light as nodes, a connection relation between the nodes is determined by a measurement feedback method or a method equivalent to the measurement feedback method; and an output layer which outputs an output signal based on the respective responses of the plurality of DOPO pulses to the input signal from the reservoir layer.
1 . An optical computer comprising:
an input layer which accepts an input signal to the optical computer;
a reservoir layer which responds to the input signal input from the input layer,
in the reservoir layer, a plurality of degenerate optical parametric oscillator (DOPO) pulses of a DOPO are generated by an optical phase-sensitive amplifier (PSA) based on pump light and are made nodes, and a connection relation between the nodes is determined by a measurement feedback method or a coupling method using optical delay lines, wherein the connection relation between the nodes is determined as a random value; and
an output layer which outputs an output signal based on the respective responses of the plurality of DOPO pulses to the input signal from the reservoir layer.
2 . The optical computer according to claim 1 ,
wherein the output signal is a signal obtained by applying a weight to each of the plurality of DOPO pulses.
3 . The optical computer according to claim 2 , further comprising:
a learning means for changing the weight by performing learning based on the output signal.
4 . The optical computer according to claim 1 ,
wherein the input signal from the input layer to the reservoir layer is a signal obtained by applying a weight to the input signal for the computer for each of the plurality of DOPO pulses.
5 . The optical computer according to claim 1 ,
wherein amplitudes of pulses of the pump light are independently determined.
6 . The optical computer according to claim 1 ,
wherein,
assuming that a cos component and a sin component of an electric field amplitude of an i-th DOPO pulse of the DOPO at an n-th cycle in a resonator are c j n and s i n respectively,
c j n and s i n are determined by following equations:
c
i
n
=
η
c
i
′
n
-
1
+
r
(
m
i
n
i
I
i
n
n
+
∑
j
J
i
j
c
j
′
n
-
1
)
s
i
n
=
η
s
i
′
n
-
1
wherein c′ i n−1 and s′ i n−1 indicate a cos component and a sin component of an electric field amplitude of the i-th DOPO pulse at an (n−1)-th cycle when being output from the PSA, l in n indicates the input signal corresponding to the i-th DOPO pulse at the n-th cycle, m in i indicates the weight corresponding to the input signal l in n , J ij indicates a matrix representing the connection relation between the i-th DOPO pulse and a j-th DOPO pulse, and η indicates a transmissivity of a DOPO optical system.
7 . The optical computer according to claim 1 ,
wherein,
assuming that a cos component of an electric field amplitude of the i-th DOPO pulse at the n-th cycle when being output from the PSA is c′ i n , an output l out n from the output layer at the n-th cycle is represented by following equation:
I
out
n
=
∑
i
m
out
i
c
i
′
n
wherein m out i indicates a weight of the output signal corresponding to the i-th DOPO pulse.