Point processes with Gaussian boson sampling.

Area of Science:

• Computational Statistics
• Quantum Information Science
• Statistical Modeling

Background:

• Random point patterns are prevalent in natural phenomena.
• Statistical point processes are essential for simulating and interpreting these patterns.
• Classical simulation of certain point processes is computationally intractable.

Purpose of the Study:

• To establish a connection between point processes and Gaussian boson sampling on photonic quantum computers.
• To develop novel quantum-inspired algorithms for simulating complex point patterns.
• To investigate the statistical properties and applications of these quantum-generated point processes.

Main Methods:

• Utilizing Gaussian boson sampling (a photonic quantum algorithm) to implement specific classes of point processes.
• Developing efficient classical algorithms inspired by quantum methods, particularly for permanental point processes.
• Analyzing the statistical properties, including clustering behavior, of the generated point patterns.

Main Results:

• Demonstrated that Gaussian boson sampling can simulate point processes based on classically intractable matrix functions.
• Introduced a family of efficient quantum-inspired point processes, including a fast classical algorithm for permanental processes.
• Revealed that these quantum-generated point processes exhibit boson-like bunching, leading to clustered point distributions.

Conclusions:

• Gaussian boson sampling offers a powerful new approach for simulating complex random point patterns.
• This work bridges quantum computing and statistical modeling, yielding efficient quantum-inspired algorithms.
• The developed methods allow for controlled generation of clustered point patterns in various state spaces.