Coherent Ising Machine Pump Light Control for Power Law Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computing devices struggle to generate and sample spin states of an Ising model according to a power law distribution, such as the Tsallis distribution, which is useful for non-equilibrium systems and applications like machine learning.
Innovation Solution
A computing device, specifically a coherent Ising machine, is designed to generate spin states of an Ising model with a distribution that can be approximated by a power law distribution, achieved by adjusting the pump light intensity to maintain the amplitude at or below the oscillation point, thereby increasing quantum fluctuation and sampling high energy states with non-negligible probability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If pump light intensity is increased to generate spin states, then the oscillation amplitude increases, but the quantum fluctuation decreases, making it difficult to sample high energy states according to power law distribution
Solution Approach 1:
The invention changes the parameter of pump light intensity to maintain the oscillation amplitude at or below the oscillation point, which increases quantum fluctuation and enables sampling of high energy states according to power law distribution. This parameter adjustment transforms the system from a regime dominated by classical oscillation to one where quantum fluctuations are significant enough to generate the desired statistical distribution.
2Reliability
If the oscillation amplitude is maintained at or below the oscillation point, then quantum fluctuation increases enabling power law distribution, but the illumination intensity is reduced
Solution Approach 1:
The invention utilizes the phase transition phenomenon at the oscillation point, where the system transitions from a stable oscillating state to a state dominated by quantum fluctuations. By operating at or below this critical point, the system undergoes a phase transition that enables the generation of spin states following power law distribution, sacrificing illumination intensity for distribution accuracy.
3Stability of the object's composition
If conventional computing devices are used with canonical distribution, then thermal equilibrium state is achieved, but non-equilibrium states with power law distribution cannot be generated
Solution Approach 1:
Instead of maintaining thermal equilibrium (canonical distribution), the invention inverts the approach by operating in a non-equilibrium regime where the oscillation amplitude is suppressed to enhance quantum fluctuations. This inversion enables the system to generate power law distributions and sample non-equilibrium states, expanding adaptability to different distribution types required for applications like machine learning.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The device effectively samples spin states according to a power law distribution, enabling applications in machine learning and other fields where such distributions are relevant, by approximating the Tsallis distribution and allowing for the sampling of high energy states.
Implementation Method 1
a pump light 3 for inducing parametric oscillation is input to the phase sensitive amplifier 2
Implementation Method 2
a phase sensitive amplifier (PSA) 2 provided in a ring-shaped optical fiber serving as a ring resonator 1 to generate a light pulse train 4
Implementation Method 3
by adjusting the pump light intensity to maintain the amplitude at or below the oscillation point, thereby increasing quantum fluctuation and sampling high energy states with non-negligible probability
Data Source
AI summary
A computing device of the present invention is a computing device for an Ising model, including a means for generating a spin state for which a distribution indicating a probability of existence of the spin state for each energy of the spin state is available to be approximated by a power law distribution.