Systems and methods for improving the performance of a quantum processor via reduced readouts
Techniques for improving the performance of a quantum processor are described. The techniques include reading out a fraction of the qubits in a quantum processor and utilizing one or more post-processing operations to reconstruct qubits of the quantum processor that are not read. The reconstructed qubits may be determined using a perfect sampler to provide results that are strictly better than reading all of the qubits directly from the quantum processor. The composite sample that includes read qubits and reconstructed qubits may be obtained faster than if all qubits of the quantum processor are read directly.
1. A method of operation in a problem solving system that comprises both a quantum processor and at least one processor-based device communicatively coupled to one another to at least approximately minimize an objective function, the quantum processor comprising a plurality of qubits including a first set of qubits and a second set of qubits, and a plurality of coupling devices, wherein each coupling device provides controllable communicative coupling between two of the plurality of qubits, the method comprising:
operating the quantum processor as a sample generator to provide samples from a probability distribution, wherein a shape of the probability distribution depends on a configuration of a number of programmable parameters for the quantum processor and a number of low-energy states of the quantum processor respectively correspond to a number of high probability samples of the probability distribution, and wherein operating the quantum processor as a sample generator comprises:
defining a configuration of the number of programmable parameters for the quantum processor via the at least one processor-based device, wherein the configuration of the number of programmable parameters corresponds to a probability distribution over the plurality of qubits of the quantum processor;
programming the quantum processor with the configuration of the number of programmable parameters via a programming subsystem;
evolving the quantum processor via an evolution subsystem; and
reading out states for the qubits in the first set of qubits of the quantum processor via a readout subsystem, wherein the states for the qubits in the first set of qubits of the quantum processor correspond to samples from the probability distribution;
processing the samples read via the readout system via the at least one processor-based device, wherein processing the samples read via the readout system via the at least one processor-based device comprises:
determining respective states for the qubits in the second set of qubits based on samples read via the readout system via the at least one processor-based device, wherein determining respective states for the qubits in the second set of qubits based on samples read via the readout system via the at least one processor-based device comprises:
sampling the states for the qubits in the second set of qubits conditioned on the states for the qubits that represent the first set of qubits read via the readout system.
2. The method of claim 1 wherein the plurality of qubits includes a third set of qubits and a fourth set of qubits, the method further comprising:
operating the quantum processor as a sample generator to provide samples from a probability distribution, wherein a shape of the probability distribution depends on a configuration of a number of programmable parameters for the quantum processor and a number of low-energy states of the quantum processor respectively correspond to a number of high probability samples of the probability distribution, and wherein operating the quantum processor as a sample generator comprises:
defining a configuration of the number of programmable parameters for the quantum processor via the at least one processor-based device, wherein the configuration of the number of programmable parameters corresponds to a probability distribution over the plurality of qubits of the quantum processor;
programming the quantum processor with the configuration of the number of programmable parameters via a programming subsystem;
evolving the quantum processor via an evolution subsystem; and
reading out states for the qubits in the third set of qubits of the quantum processor via a readout subsystem, wherein the states for the qubits in the third set of qubits of the quantum processor correspond to samples from the probability distribution;
processing the samples read via the readout system via the at least one processor-based device, wherein processing the samples read via the readout system via the at least one processor-based device comprises:
determining respective states for the qubits in the fourth set of qubits based on samples read via the readout system via the at least one processor-based device.
3. The method of claim 1 wherein processing the samples read via the readout system via the at least one processor-based device comprises processing the samples read via the readout system via at least one of a microprocessor, a digital signal processor (DSP), a graphical processing unit (GPU), or a field programmable gate array (FPGA).
4. The method of claim 1 wherein determining respective states for the qubits in the second set of qubits based on samples read via the readout system via the at least one processor-based device comprises executing at least one of: an optimization operation, an enumeration, a sampling operation or evaluation of estimators.
5. The method of claim 1 wherein determining respective states for the qubits in the second set of qubits based on samples read via the readout system via the at least one processor-based device comprises executing at least one of: a local gradient descent procedure or a Gibbs sampling procedure.
6. The method of claim 1 , further comprising:
selectively modifying the first set of qubits and the second set of qubits to change which ones of the qubits of the quantum processor constitute the first set of qubits and which ones of the qubits of the quantum processor constitute the second set of qubits.
7. The method of claim 1 wherein determining respective states for the qubits in the second set of qubits based on samples read via the readout system via the at least one processor-based device includes performing a classical heuristic optimization algorithm to determine states for the qubits in the second set of qubits based on samples read via the readout system via the at least one processor-based device.
8. The method of claim 7 wherein performing a classical heuristic optimization algorithm to determine states for the qubits in the second set of qubits based on samples read via the readout system via the at least one processor-based device includes performing at least one of: a majority voting on chains of qubits post-processing operation, a local search to find a local minima post-processing operation, or a Markov Chain Monte Carlo simulation at a fixed temperature post-processing operation.
9. The method of claim 1 wherein evolving the quantum processor via an evolution subsystem includes performing at least one of adiabatic quantum computation or quantum annealing.
10. The method of claim 1 wherein operating the quantum processor as a sample generator comprises:
reading out states for the qubits in the second set of qubits of the quantum processor via the readout subsystem, wherein the states for the qubits in the first set of qubits of the quantum processor correspond to samples from the probability distribution;
wherein processing the samples read via the readout system via the at least one processor-based device comprises:
determining respective states for the qubits in the first set of qubits based on the samples read via the readout system via the at least one processor-based device.