# Techniques: Improved Accreditation of Analogue Quantum Simulation and Establishing Quantum Advantage

## What the paper contributes (one paragraph)
The paper presents an improved protocol for accrediting the outputs of analogue (and hybrid digital-analogue) quantum simulations. Given a target time evolution, it runs the requested simulation alongside a set of trap simulations that are equivalent to the identity; counting how often the traps return the correct outcome yields a rigorous upper bound on the ideal-actual variation distance of the target run, with chosen accuracy and confidence. The central enabler is a method to approximately invert any spin Hamiltonian using only added single-qubit gates, removing the earlier need for universal Hamiltonians, extra qubits, and two-qubit gates, and dropping the assumption that error is identical across runs. Because the bound is exactly the metric used by complexity-theoretic results, the protocol gives a near-term route to verifying a quantum advantage demonstration based on analogue simulation.

## Techniques offered
- **Analogue quantum simulation accreditation** - a trap-based protocol that runs the target simulation among identity-equivalent traps and returns a certified upper bound on the simulation's ideal-actual (total variation) distance.
  - guarantee: statistical bound; an explicit upper bound on the ideal-actual variation distance with accuracy theta and confidence alpha, using a stated number of trap simulations derived from Hoeffding's inequality.
  - quote: "It returns the same measurement outcomes as an analogue quantum simulation, and additionally an $\epsilon_{\text{VD}} \in [0,1)$ that upper bounds the ideal-actual variation distance (in Def. \ref{def:idealActualVarDist}) of the execution from which the measurement outcomes are obtained, with accuracy $\theta$ and confidence $\alpha$."
- **Approximate Hamiltonian time inversion via single-qubit gates** - a compilation method that turns a forward evolution of any spin Hamiltonian into an approximate reverse evolution purely by inserting random single-qubit gates, enabling the identity-equivalent trap construction without universal Hamiltonians or two-qubit gates.
  - guarantee: static/constructive bound; the constructed sub-circuit approximates the inverse evolution to arbitrary additive error in the diamond norm, with explicit resource scaling.
  - quote: "For any $N$-qubit spin Hamiltonian, $\mathcal{H}$, and time duration, $t \in \mathbb{R}$, there exists a sub-circuit (that we denote $\mathcal{B} (\mathcal{H}, t, \epsilon, 0)$) implementing $e^{-i \mathcal{H}t}$ that, purely by adding single-qubit gates, can be transformed into a sub-circuit (that we denote $\mathcal{B} (\mathcal{H}, t, \epsilon, 1)$) which approximately implements $e^{i \mathcal{H}t / (L-1)}$ to arbitrary additive error in the diamond norm, $\epsilon \in \mathbb{R}$ , where $L \in \mathbb{N}$ is efficiently computable."
- **Verification for simulation-based quantum advantage** - the protocol supplies exactly the ideal-actual variation distance figure that complexity-theoretic advantage proposals require, so it can certify an analogue simulation that is classically intractable below a known threshold.
  - guarantee: statistical bound feeding a complexity-theoretic claim; certifies the variation distance lies below the threshold (about 0.292) under which the simulation is classically intractable given stated assumptions.
  - quote: "That the requirement for quantum advantage in the experiment proposed in Ref.~\cite{PhysRevX.8.021010} is expressed in terms of the ideal-actual variation distance additionally makes the analogue accreditation protocol presented herein well suited for use in verifying it, as our protocol returns exactly this metric about the quality of a simulation's execution."

## Where it could apply
- Target primitive(s): Trotterised and analogue Hamiltonian simulation (time evolution); sampling-based quantum advantage experiments.
- Target application group(s): platforms_tooling, chemistry_materials, error_correction (verification/accreditation), and quantum advantage demonstration.
- Code family / hardware assumptions: programmable hybrid analogue-digital quantum simulators with single-qubit gates but no required two-qubit gates; explicitly suited to neutral-atom platforms; no extra qubits required; spin Hamiltonians.

## Caveats
none