Abstract:It is shown that the gap that limits the speed of a quantum annealing process can take three salient morphologies: (i) the supersteep gap where both the ground and the first excited eigenenergy level curves have topologically related pairs of nearby inflection points giving both the maximum and the minimum a steep aspect, (ii) the steep gap where only the first excited eigenenergy level curve has a pair of inflection points giving its minimum a steep aspect while the maximum of the ground level does not exhibit inflection points, and (iii) the mild gap that has no related inflection points. Classification of the various singularities betrayed by the inflection points relies on the critical value curves of the quadratic numerical range mapping of the matrix H0+iH1, where H0 is the transverse field Hamiltonian and H1 the problem Hamiltonian. It is shown that the ground level is mapped to the generically smooth boundary of the numerical range, while the first excited level is mapped to an interior non-smooth critical value curve exhibiting swallow tails. The major result is that the position of the swallow tails relative to the boundary allows the supersteep versus steep discrimination, while the absence of swallow tail-boundary interaction characterizes the mild gap. As a corollary of the singularity analysis, the highly structured initial and final Hamiltonians of the Grover search create unstable singularities that break into stable swallow tails under perturbation, with the consequence of invalidating the gap scaling estimates computed around the unstable singularity. Classification of all stable singularities from a global viewpoint requires the Legendrian approach where the energy level curves become Legendrian knots in the contact space. Last but not least, it will be shown that a supersteep swallow tail previews tunneling.

Spectral Gap Analysis for Efficient Tunneling in Quantum Adiabatic Optimization

Discrepancies between Asymptotic and Exact Spectral Gap Analyses of Quantum Adiabatic Barrier Tunneling

Quantum Monte Carlo Simulations of Tunneling in Quantum Adiabatic Optimization

Necessary Adiabatic Run Times in Quantum Optimization

Conditional t-independent spectral gap for random quantum circuits and implications for t-design depths

Topological Discrimination of Steep to Supersteep Gap as Evidence of Tunneling in Adiabatic Quantum Processes

Unstructured Adiabatic Quantum Optimization: Optimality with Limitations

Spectral Gap Optimization for Enhanced Adiabatic State Preparation

Bounding first-order quantum phase transitions in adiabatic quantum computing

Circumventing superexponential runtimes for hard instances of quantum adiabatic optimization

Lower bounds for adiabatic quantum algorithms by quantum speed limits

Quantum Adiabatic Evolutions That Can't Be Used to Design Efficient Algorithms

Computational Role of Collective Tunneling in a Quantum Annealer

Staying adiabatic with unknown energy gap

Space-efficient binary optimization for variational quantum computing

Spectral Gap Superposition States

Universal scaling of adiabatic tunneling out of a shallow confinement potential

Thermodynamic and energetic constraints on out-of-equilibrium tunneling rates

On Quantum Speedups for Nonconvex Optimization Via Quantum Tunneling Walks

Gates for Adiabatic Quantum Computing

Fighting Exponentially Small Gaps by Counterdiabatic Driving