# Techniques: Adiabatic quantum computing with parameterized quantum circuits

## What the paper contributes (one paragraph)
The paper proposes AQC-PQC, a hybrid quantum-classical ground-state preparation algorithm that mimics adiabatic quantum computing on a fixed parameterized quantum circuit without running an energy minimisation at each step. The central result is a perturbation-theory theorem: given the angles that minimise the energy of a known Hamiltonian, a small Hamiltonian perturbation shifts those angles by a vector that solves a constrained linear system, whose coefficients (a gradient vector and a Hessian) are obtained by measuring observables and their derivatives on the unperturbed state via finite differences or parameter-shift rules. Applied iteratively along a discretised adiabatic path from an easy Hamiltonian to a target Hamiltonian, this tracks the ground state in a fixed number of steps K. The authors show the per-step problem can be cast as a semidefinite program, quantify the quantum resources required, and benchmark against VQE on MaxCut, Number Partitioning, and the transverse-field Ising chain.

## Techniques offered
- **AQC-PQC ground-state preparation (predictor-corrector / discrete adiabatic eigensolver on a parameterized circuit)** - tracks the ground state of a target Hamiltonian within a fixed parameterized-circuit ansatz by analytically computing parameter shifts from measured observables along a discretised adiabatic path, replacing per-step variational optimisation and so avoiding random-initialisation and barren-plateau failure modes.
  - guarantee: provided the evolution is slow enough and the ansatz expressive enough, the method is guaranteed to find a good approximation to the ground state; per-step quantum resources are bounded and cast as a semidefinite program, but the assurance is conditional on the spectral gap and degrades under noise.
  - quote: "We formulate an algorithm to find the best approximation of the ground state of a Hamiltonian within a family of parameterized quantum states that: (i) can be applied in an early fault-tolerant device, (ii) is not sensitive to the initialization points and (iii) requires fixed calls to the quantum computer with theoretical guarantees on the performance."

## Where it could apply
- Target primitive(s): VQE, QAOA, adiabatic / Trotterised simulation, parameterized quantum circuits with parameter-shift rules
- Target application group(s): optimization, chemistry_materials, quantum_ml, platforms_tooling

## Caveats
This is principally a new-algorithm paper; the suppliable item is the AQC-PQC method itself as a drop-in alternative to a vendor's VQE-style eigensolver, not a verification, decoding, or mitigation primitive layered onto an external result. The perturbation theorem, the semidefinite-program reformulation, and the O((1 + dim N_kappa(A)) M^2) per-step quantum-resource count are guarantees or internal sub-parts of that one algorithm rather than separable techniques. Guarantees are noiseless and conditional on the spectral gap and ansatz expressivity; robustness to noise and the use of error mitigation are flagged as future work. Results are small simulated instances only. Open access (arXiv), no paywall.