This disclosure relates to evaluating and improving performance of a control implementation on a quantum processor comprising multiple qubits in the presence of noise. A noise model decomposes noise interactions described by a multi-qubit noise Hamiltonian into multiple contributory noise channels. Each channel generates noise dynamics described by a unique noise-axis operator. For a given control implementation, a unique filter function represents susceptibility of the multi-qubit system to the associated noise dynamics. The filter functions are based on a frequency transformation of the noise axis operator of the corresponding noise channel to thereby evaluate the performance of the control implementation. An optimised control sequence is based on the filter function to reduce the susceptibility of the multi-qubit system to the noise channels, thereby reducing the effective interaction with the multi-qubit noise Hamiltonian. The optimised control sequence controls the quantum processor to thereby improve the performance of the control implementation.
Claim CLM-00001. 1. A method for evaluating and improving performance of a control implementation on a quantum processor comprising multiple qubits in the presence of noise, the method comprising:
modelling the noise by decomposing noise interactions described by a multi-qubit noise Hamiltonian into multiple contributory noise channels, each channel generating noise dynamics described by a unique noise-axis operator A{circumflex over ( )}{(i)}, where (i) indexes the ith noise channel; determining, for a given control implementation, the unique filter function for each noise channel representing susceptibility of the multi-qubit system to the associated noise dynamics, each of the multiple filter functions being based on a frequency transformation ($\mathcal{F}$) of the noise axis operator (A{circumflex over ( )}{(i)}) of the corresponding noise channel (i) to thereby evaluate the performance of the control implementation; determining an optimised control sequence based on the filter function to reduce the susceptibility of the multi-qubit system to the noise channels, thereby reducing the effective interaction with the multi-qubit noise Hamiltonian; and applying the optimised control sequence to the multi-qubit system to control the quantum processor to thereby improve the performance of the control implementation.
Claim CLM-00019. 19. A quantum processor comprising multiple qubits and a controller configured to:
model the noise by decomposing noise interactions described by a multi-qubit noise Hamiltonian into multiple contributory noise channels, each channel generating noise dynamics described by a unique noise-axis operator A{circumflex over ( )}{(i)}, where (i) indexes the ith noise channel; determine, for a given control implementation, the unique filter function for each noise channel representing susceptibility of the multi-qubit system to the associated noise dynamics, each of the multiple filter functions being based on a frequency transformation ($\mathcal{F}$) of the noise axis operator (A{circumflex over ( )}{(i)}) of the corresponding noise channel (i) to thereby evaluate the performance of the control implementation; determine an optimised control sequence based on the filter function to reduce the susceptibility of the multi-qubit system to the noise channels, thereby reducing the effective interaction with the multi-qubit noise Hamiltonian; and apply the optimised control sequence to the multi-qubit system to control the quantum processor to thereby improve the performance of the control implementation.
