One of the main methods for protecting quantum information against decoherence is to encode information in the ground subspace (or the low energy sector) of a Hamiltonian with a large energy gap which penalizes errors from environment. The protecting Hamiltonian is chosen such that its degenerate ground subspace is an error detecting code for the errors caused by the interaction with the environment. We consider environments with arbitrary number of local sites, e.g. spins, whose interactions among themselves are local and bounded. Then, assuming the system is interacting with a finite number of sites in the environment, we prove that, up to second order with respect to the coupling constant, decoherence and relaxation are suppressed by a factor which grows exponentially fast with the ratio of energy penalty to the norm of local interactions in the environment. The state may, however, still evolve unitarily inside the code subspace due to the Lamb shift effect. In the context of adiabatic quantum computation, this means that the evolution inside the code subspace is effectively governed by a renormalized Hamiltonian. The result is derived from first principles, without use of master equations or their assumptions, and holds even in the infinite temperature limit. We also prove that unbounded or non-local interactions in the environment at sites far from the system do not considerably modify the exponential suppression. Our main technical tool is a new bound on the decay of power spectral density at high frequencies for local observables and many-body Hamiltonians with bounded and local interactions in a neighborhood around the support of the observable.
One of the main methods for protecting quantum information against decoherence is to encode information in the ground subspace (or the low energy sector) of a Hamiltonian with a large energy gap which penalizes errors from environment. The protecting Hamiltonian is chosen such that its degenerate ground subspace is an error detecting code for the errors caused by the interaction with the environment. We consider environments with arbitrary number of local sites, e.g. spins, whose interactions among themselves are local and bounded. Then, assuming the system is interacting with a finite number of sites in the environment, we prove that, up to second order with respect to the coupling constant, decoherence and relaxation are suppressed by a factor which grows exponentially fast with the ratio of energy penalty to the norm of local interactions in the environment. The state may, however, still evolve unitarily inside the code subspace due to the Lamb shift effect. In the context of adiabatic quantum computation, this means that the evolution inside the code subspace is effectively governed by a renormalized Hamiltonian. The result is derived from first principles, without use of master equations or their assumptions, and holds even in the infinite temperature limit. We also prove that unbounded or non-local interactions in the environment at sites far from the system do not considerably modify the exponential suppression. Our main technical tool is a new bound on the decay of power spectral density at high frequencies for local observables and many-body Hamiltonians with bounded and local interactions in a neighborhood around the support of the observable.
Hamiltonian quantum computing, such as the adiabatic and holonomic models, can be protected against decoherence using an encoding into stabilizer subspace codes for error detection and the addition of energy penalty terms. This method has been widely studied since it was first introduced by Jordan, Farhi, and Shor (JFS) in the context of adiabatic quantum computing. Here, we extend the original result to general Markovian environments, not necessarily in Lindblad form. We show that the main conclusion of the original JFS study holds under these general circumstances: Assuming a physically reasonable bath model, it is possible to suppress the initial decay out of the encoded ground state with an energy penalty strength that grows only logarithmically in the system size, at a fixed temperature.
We consider the use of quantum error-detecting codes, together with energy penalties against leaving the code space, as a method for suppressing environmentally induced errors in Hamiltonian-based quantum computation. This method was introduced in Jordan et al. [Phys. Rev. A 74, 052322 (2006)] in the context of quantum adiabatic computation, but we consider it more generally. Specifically, we consider a computational Hamiltonian, which has been encoded using the logical qubits of a single-qubit error-detecting code, coupled to an environment of qubits by interaction terms that act one-locally on the system. Additional energy penalty terms penalize states outside of the code space. We prove that in the limit of infinitely large penalties, one-local errors are completely suppressed, and we derive some bounds for the finite penalty case. Our proof technique involves exact integration of the Schrodinger equation, making no use of master equations or their assumptions. We perform long time numerical simulations on a small (one logical qubit) computational system coupled to an environment and the results suggest that the energy penalty method achieves even greater protection than our bounds indicate.
Adiabatic quantum computing (AQC) can be protected against thermal excitations via an encoding into error-detecting codes, supplemented with an energy penalty formed from a sum of commuting Hamiltonian terms. Earlier work showed that it is possible to suppress the initial thermally induced excitation out of the encoded ground state, in the case of local Markovian environments, by using an energy-penalty strength that grows only logarithmically in system size, at a fixed temperature. The question of whether this result applies beyond the initial time was left open. Here we answer this in the affirmative. We show that thermal excitations out of the encoded ground state can be suppressed at arbitrary times under the additional assumption that the total evolution time is polynomial in the system size. Thus, computational problems that can be solved efficiently by using AQC in a closed system setting can still be solved efficiently subject to coupling to a thermal environment. Our construction uses stabilizer subspace codes, which require at least 4-local interactions to achieve this result.
We consider error suppression schemes in which quantum information is encoded into the ground subspace of a Hamiltonian comprising a sum of commuting terms. Since such Hamiltonians are gapped, they are considered natural candidates for protection of quantum information and topological or adiabatic quantum computation. However, we prove that they cannot be used to this end in the two-local case. By making the favorable assumption that the gap is infinite, we show that single-site perturbations can generate a degeneracy splitting in the ground subspace of this type of Hamiltonian which is of the same order as the magnitude of the perturbation, and is independent of the number of interacting sites and their Hilbert space dimensions, just as in the absence of the protecting Hamiltonian. This splitting results in decoherence of the ground subspace, and we demonstrate that for natural noise models the coherence time is proportional to the inverse of the degeneracy splitting. Our proof involves a new version of the no-hiding theorem which shows that quantum information cannot be approximately hidden in the correlations between two quantum systems. The main reason that two-local commuting Hamiltonians cannot be used for quantum error suppression is that their ground subspaces have only short-range (two-body) entanglement.
Near-Optimal Bounds on the Density of Low-Energy States of k-Local Hamiltonians and Faster Quantum Algorithms
Low-energy estimation and state preparation for general k-local Hamiltonians are fundamental challenges in quantum complexity theory. Buhrman et al.~ [BGLGST, PRL 2025] recently broke the natural Grover bound O∗(2n/2) for both problems, with the improvement depending on the relative accuracy ε and the locality k. In this work, we present faster exponential quantum algorithms for these problems, where the binary entropy function governs the runtime exponent. For sufficiently small ε/k, our algorithms improve the exponent by a factor of log(k/ε) over [BGLGST, PRL 2025]. Our main technical result is an entropy-governed lower bound on the dimension of the Hamiltonian's low-energy subspace, obtained by depolarizing its ground state. For fixed k, this bound is optimal up to constant factors in the exponent. The same framework yields tighter bounds for Heisenberg, XY, and Ising models on arbitrary interaction graphs.
Achieving noise resilience is an outstanding challenge in Hamiltonian-based quantum computation. To this end, energy-gap protection provides a promising approach, where the desired quantum dynamics are encoded into the ground space of a penalty Hamiltonian that suppresses unwanted noise processes. However, existing approaches either explicitly require high-weight penalty terms that are not directly accessible in current hardware, or utilize non-commuting 2-local Hamiltonians, which typically leads to an exponentially small energy gap. In this work, we provide a general recipe for designing error-resilient Hamiltonian simulations, making use of an excited encoding subspace stabilized by solely 2-local commuting Hamiltonians. Our results thus overcome a no-go theorem previously derived for ground-space encoding that prevents noise suppression schemes with such Hamiltonians. Importantly, our method is scalable as it only requires penalty terms that scale polynomially with system size. To illustrate the utility of our approach, we further apply this method to a variety of 1- and 2-dimensional many-body spin models, potentially extending the duration of high-fidelity simulation by orders of magnitude in current hardware.