Multidimensional hyperspin machine

Area of Science:

• Physics
• Computer Science
• Quantum Computing

Background:

• Multidimensional spins are crucial in fields from condensed matter physics to quantum chromodynamics.
• Current machines simulating spin models are limited to low-dimensional or Ising spins.
• Simulating arbitrary-dimensional spin models remains a significant computational challenge.

Purpose of the Study:

• To introduce and validate a novel hyperspin machine capable of simulating multidimensional continuous spin models.
• To demonstrate the machine's ability to find approximate ground states for complex graphs.
• To present dimensional annealing as a strategy for interpolating between spin model dimensions.

Main Methods:

• Realizing high-dimensional spins through pumping groups of parametric oscillators.
• Utilizing dimensional annealing by tuning coupling topology to interpolate between dimensions.
• Comparing performance against conventional Ising simulators for specific model interpolations.

Main Results:

• The hyperspin machine accurately approximates the ground state of complex graphs.
• Dimensional annealing significantly boosts success probability when interpolating between XY and Ising models.
• The hyperspin machine demonstrates a new computational paradigm for combinatorial optimization.

Conclusions:

• Hyperspin machines offer a viable approach for simulating complex, multidimensional spin models.
• This technology can be realized with off-the-shelf hardware for diverse applications.
• The hyperspin machine opens new avenues in classical and quantum computing, and fundamental physics research.