Simulating physical models is an ancient tool of science, from mechanical planetariums to differential analyzers. Today, many important problems, from high-temperature superconductivity to molecular energies, hit the limits of classical computation. Quantum devices promise to break these barriers, but the community has not yet reached a consensus on what counts as a quantum simulator and how to trust it. This work joins the ranks of reviews and conceptual articles on quantum simulation, but focuses on fundamental definitions and epistemology rather than surveying experimental platforms.
The authors conduct a conceptual analysis, drawing on existing definitions and examples from the literature. They introduce a formal criterion for classicality: a device is classical if, in some single-particle basis, the density matrix is diagonal at all times. They then apply this criterion to various devices, from semiconductor chips to Bose condensates, and discuss how to classify analog and digital simulators, as well as quantum annealing. To justify the need for quantum simulators, they list the limitations of classical numerical methods: exact diagonalization, mean-field methods, tensor networks, density functional theory, and quantum Monte Carlo.
The main result is a clear distinction between simulation and computation. Simulation is the first step of a two-step process: first, the device reproduces a model; then, the model is compared with the real system. Computation, in contrast, implies guaranteed accuracy and does not require reference to reality. On this basis, Lloyd's digital quantum simulator with error correction should be considered a computer, and the D-Wave device for finding the ground state of the Ising model is a computer, not a simulator, because its goal is optimization, not studying a physical system. Classicality is determined by the diagonality of the density matrix in a single-particle basis, which allows even a Bose condensate to be classified as a classical simulator if it operates in the condensate regime, but as quantum when using Bogoliubov excitations. The authors stress that trust in a simulator is not an objective property but the result of a falsification process: tests, comparisons, and gradual accumulation of confidence.
This work differs from neighboring reviews by shifting the focus from enumerating platforms and achievements to the philosophy and methodology of quantum simulation. While many articles describe how to build simulators, this one asks what counts as a simulator and when it can be trusted. It is neither a new experiment nor a new algorithm, but an attempt to bring order to concepts, which is critical as we move into regimes inaccessible to classical verification.
As quantum simulators begin to outperform classical computers on problems with no exact solutions, the verification problem will become central. The authors predict that efforts will shift from demonstrating control to developing methods for establishing trust: finding regimes where results are sensitive to errors, and understanding which quantities can be reliably simulated at all. New approaches to falsification may emerge, such as comparing different simulators or using symmetries.
This article will influence the quantum simulator community by providing a common language and evaluation criteria. It may also be useful to philosophers of science and quantum computing specialists, clarifying the boundaries between simulation and computation.
Practical protocols must be developed for verifying analog quantum simulators in regimes where classical methods fail. It is also important to investigate which physical quantities are robust to model errors and which are not.
An open question in quantum physics is how to distinguish truly quantum behavior from effectively classical behavior in many-body systems. This work is directly related to the problem of defining quantumness and to the difficulty of simulating strongly correlated systems, where classical methods fail due to the exponential growth of Hilbert space or the sign problem.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The authors carry out a conceptual analysis, introducing a formal criterion for classicality (diagonality of the density matrix in a single-particle basis) and applying it to various devices. They distinguish between simulation and computation, classify analog and digital simulators, as well as quantum annealing, and discuss how to establish trust in simulators through falsification.
Where its strength is
A clear distinction between simulation and computation based on purpose and relation to reality, along with an operational criterion for classicality that can be applied to a wide class of devices.
Where it could go next
Worth trying
Why this may matter
Refining definitions and trust criteria is important as we move into regimes where classical verification is impossible; this could influence standards and practices in the quantum technology community.
Nearby works in our archive
The archive is crowded around this work: 2 reviews nearby, closest at 0.72. This ground is well walked — and there is plenty to compare against.
The term 'quantum simulator' is often attributed to Richard Feynman, who in 1982 suggested that quantum systems are best modeled by quantum devices. However, the idea of using one physical system to study another dates back to ancient mechanical models of celestial bodies.