Multiphoton Quantum Simulation of the Generalized Hopfield Memory Model

Area of Science:

• Quantum optics and artificial intelligence.
• Explores the intersection of quantum interference and classical neural network models.

Background:

• Multiphoton quantum interference is key to photonic quantum technologies.
• Hopfield-like Hamiltonians model associative memory and machine learning.

Purpose of the Study:

• To establish a connection between multiphoton quantum interference and Hopfield-like Hamiltonians.
• To investigate the use of photonic systems as quantum simulators for classical models.

Main Methods:

• Utilized a system of indistinguishable photons, phase shifters, and a linear-optical interferometer.
• Mapped output photon statistics to a p-body Hopfield Hamiltonian (p=2N_{ph}).
• Investigated a generalized four-body Hopfield model.

Main Results:

• Demonstrated that photonic systems can generate output statistics described by Hopfield Hamiltonians.
• Observed a transition from memory retrieval to a spin-glass phase in the simulated model.
• Showcased the potential for photonic quantum simulators to study complex classical systems.

Conclusions:

• The developed mapping provides novel routes for simulating disordered classical systems using photonic quantum simulators.
• This work bridges quantum technologies and classical artificial intelligence, offering new perspectives on both fields.