Quantum control link distortion correction optimization method and device and quantum computer

Abstract

The application discloses a quantum control link distortion correction optimization method and device and a quantum computer, and belongs to the technical field of quantum computing. A quantum control link to be corrected is connected with a target quantum bit. The method comprises the following steps: acquiring a seed library comprising a plurality of seeds; wherein each seed represents a pole model parameter of a digital filter used for performing distortion correction on a corrected quantum control link, and the digital filter is determined based on a system function fusing quantum control link signal response characteristics; performing an optimization algorithm based on the pole model parameters of the seeds as initial values of to-be-optimized parameters, and applying a frequency control signal on a Z link of the target quantum bit according to the optimized pole model parameters to obtain the fidelity of the quantum state of the target quantum bit; and determining the pole model parameters when the fidelity meets a preset threshold as target correction parameters of the quantum control link to be corrected. The application can greatly improve the distortion optimization efficiency of the Z link.

Description

Technical Field

[0001] This application relates to the field of quantum computing, and in particular to a distortion correction optimization method, device, and quantum computer for a quantum measurement and control link. Background Technology

[0002] A quantum computer system comprises a quantum chip system (also called a quantum processor), a quantum computer control system, a quantum computer environment support system (also called a dilution refrigerator), and a quantum computer operating system integrated into a classical computer. The dilution refrigerator provides different temperature zones ranging from room temperature to the lowest temperature using cold plates, which isolate these zones. The quantum chip system is located in the lowest temperature zone of the dilution refrigerator. The quantum computer control system connects to and controls the quantum chip system via a quantum control link. This link consists of a room temperature link, a room temperature-low temperature link (also called a cross-temperature zone link), and a low temperature link (also called an ultra-low temperature link), connected sequentially. The room temperature link connects the quantum computer control system to the top flange of the dilution refrigerator. The room temperature-low temperature link provides signal connections between the cold plates corresponding to the highest and lowest temperature zones within the dilution refrigerator, while the low temperature link provides signal connections between the cold plate corresponding to the lowest temperature zone and the quantum processor within the lowest temperature zone.

[0003] In transmitting quantum signals, quantum telemetry and control links not only cause transmission delays but also signal distortion. This is because the coaxial cable constituting the quantum telemetry and control link and the devices mounted on it are not ideal. The distortions that occur during the generation and transmission of quantum signals result in waveforms that ultimately reach the qubits that are distorted compared to the designed waveforms, which severely affects the fidelity of operations in quantum computing and quantum simulation. Therefore, distortion correction for quantum telemetry and control links is necessary. Summary of the Invention

[0004] The purpose of this application is to provide a distortion correction optimization method, device, and quantum computer for quantum measurement and control links, so as to overcome the shortcomings of the prior art and enable distortion correction to be performed quickly and effectively.

[0005] To address the aforementioned technical problems, the first aspect of this application provides a distortion correction optimization method for a quantum measurement and control link, wherein the quantum measurement and control link to be corrected is connected to a target qubit, and the method includes:

[0006] A seed library containing multiple seeds is obtained; wherein each seed represents the pole model parameters of a digital filter used to perform distortion correction on a calibrated quantum measurement and control link, the digital filter being determined based on a system function that incorporates the signal response characteristics of the quantum measurement and control link; The optimization algorithm is executed based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, and the frequency control signal is applied on the Z link of the target qubit according to the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit. The pole model parameters that satisfy the preset threshold are determined as the target calibration parameters of the quantum measurement and control link to be calibrated.

[0007] In the distortion correction optimization method for the quantum measurement and control link described above, optionally, the system function includes pure zero-point terms, real pole terms, and complex pole terms, and the pole model parameters of each seed include several real pole term parameters and complex pole term parameters; wherein, the pure zero-point terms characterize the finite-length impulse response of the quantum measurement and control link, the real pole terms characterize the exponential response of the quantum measurement and control link, and the complex pole terms characterize the oscillation decay response of the quantum measurement and control link.

[0008] The distortion correction optimization method for the quantum measurement and control link described above may optionally include the Nelder-Mead algorithm or the differential evolution algorithm.

[0009] The distortion correction optimization method for the quantum telemetry and control link described above, optionally, when the optimization algorithm includes a differential optimization algorithm, before executing the optimization algorithm based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, the method further includes: Determine the target population size for the differential evolution algorithm; Based on the pole model parameters of all seeds in the seed library, the fidelity test of the quantum state of the qubit is performed on each quantum measurement and control link; The first seed bank is determined by sorting the seeds according to the size of the fidelity test results to determine the target population size mentioned above. The optimization algorithm is executed using the pole model parameters based on several seeds as initial values ​​for the parameters to be optimized, including: The optimization algorithm is executed based on the pole model parameters of several seeds in the first seed library as the initial values ​​of the parameters to be optimized.

[0010] The distortion correction optimization method for the quantum measurement and control link described above, optionally, includes applying a frequency control signal to the Z-link of the target qubit based on the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit, comprising: A frequency control signal for loading a predistortion waveform is applied to the Z-link of the target qubit according to the optimized pole model parameters; wherein the predistortion waveform is generated by the digital filter based on the pole model parameters; A logic gate signal for quantum state operation is applied to the XY control line of the target qubit; Obtain the fidelity of the quantum state output by the target qubit.

[0011] In the distortion correction optimization method for the quantum telemetry and control link described above, optionally, the digital filter generates the pre-distortion waveform based on the pole model parameters, including: The coefficients of the digital filter are determined based on the pole model parameters. The digital filter generates a predistorted waveform based on the coefficients.

[0012] In the distortion correction optimization method of the quantum measurement and control link described above, optionally, the interval between the frequency control signal and the logic gate signal is set so that the impulse response of finite length has no effect on the quantum state change of the target quantum bit.

[0013] Optionally, the distortion correction optimization method for the quantum telemetry and control link described above may further include: The target correction parameters are used as the new seed to update the seed library.

[0014] A second aspect of this application provides a distortion correction and optimization device for a quantum measurement and control link, wherein the quantum measurement and control link to be corrected is connected to a target qubit, and the device includes: A seed acquisition module is used to acquire a seed library containing multiple seeds; wherein each seed represents the pole model parameters of a digital filter used to perform distortion correction on a calibrated quantum measurement and control link, and the digital filter is determined based on a system function that integrates the signal response characteristics of the quantum measurement and control link; The algorithm execution module is used to execute the optimization algorithm based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, and to apply a frequency control signal on the Z link of the target qubit according to the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit. The parameter determination module is used to determine the pole model parameters when the fidelity meets a preset threshold as the target correction parameters of the quantum measurement and control link to be corrected.

[0015] A third aspect of this application provides a quantum computer that uses a distortion correction optimization method for a quantum measurement and control link as described in any of the first aspects above to correct the quantum measurement and control link connecting the qubits, or includes a distortion correction optimization device for a quantum measurement and control link as described in the second aspect above.

[0016] Compared with existing technologies, this application obtains a seed library and reuses the pole model parameters of the digital filters used in the Z-link that has previously undergone distortion correction on new, similar links. This avoids the need to collect and fit the time-domain response from scratch, which not only greatly simplifies the testing process and saves a lot of testing time, but also provides a starting point closer to the global optimum for the optimization algorithm because the seed parameters come from real optimization data results with a large sample size and higher reliability. Furthermore, as the seed library gradually expands, the initial population and the actual parameters of the Z-link to be tested are more likely to coincide, which can significantly increase optimization efficiency and convergence speed.

[0017] The distortion correction optimization device and quantum computer for the quantum measurement and control link provided in this application belong to the same concept as the distortion correction optimization method for the quantum measurement and control link, and therefore have the same beneficial effects, which will not be elaborated here. Attached Figure Description

[0018] Figure 1 A flowchart illustrating a distortion correction optimization method for a quantum measurement and control link provided in an embodiment of this application; Figure 2 A schematic diagram of a unit step response of a multi-link telemetry and control system provided in this application embodiment; Figure 3 An example diagram of a seed bank is provided for an embodiment of this application; Figure 4 This application provides a schematic diagram of the experimental principle for performing random benchmark tests in an embodiment of the present application; Figure 5 This application provides a schematic diagram illustrating the results of performing random benchmark tests under different conditions. Figure 6 This is a schematic diagram of the structure of a distortion correction and optimization device for a quantum measurement and control link provided in an embodiment of this application. Detailed Implementation

[0019] The specific embodiments of this application will be described in more detail below with reference to the schematic diagrams. The advantages and features of this application will become clearer from the following description and claims. It should be noted that the drawings are all in a very simplified form and use non-precise proportions, and are only used to facilitate and clarify the illustration of the embodiments of this application.

[0020] In the description of this application, it should be understood that the terms "center", "upper", "lower", "left", "right", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0022] Quantum measurement and control links include XY and Z links. When these links transmit quantum signals, they not only cause transmission delays but also distortions. This is because the coaxial cables and devices mounted on them are not ideal. During the research process, the applicant summarized the factors that may cause quantum signal distortion as follows: 1. Imperfect signal generation: The generation process of quantum signals involves generating control pulses required by an arbitrary waveform generator. Arbitrary waveform generators have limited bandwidth and may have overshoot. This results in the waveform output from the arbitrary waveform generator deviating from the original design waveform. 2. Various devices are installed on the quantum measurement and control links, such as filters, DC blockers, and bias tees, each with its own frequency response. The skin effect of the coaxial cable also leads to a longer rise time in the system response. 3. Reflections caused by impedance mismatch, especially when the quantum chip is mounted on the package, due to impedance control of the wiring bonding line.

[0023] Distortions occurring during the generation and transmission of quantum signals can cause the waveform arriving at the qubit to deviate from the original design, severely impacting the fidelity of operations in quantum computing and simulation. For example, two-qubit gates in frequency-tunable qubit systems are typically implemented by applying a pulse of tens of nanoseconds to the Z-line to adjust the bit frequency. Quantum simulations also evolve by rapidly adjusting the bit to a specific frequency using Z-pulses. Specifically, pulse tailing caused by Z-line distortion affects the bit frequency when a single-qubit gate is subsequently applied, resulting in a decrease in the fidelity of the corrected single-qubit gate. For two-qubit gates, such as control phase gates, short-timescale distortions can cause the bit evolution to deviate from its carefully designed trajectory, leading to control phase errors and the eventual leakage of the bit state into non-computational subspaces. More serious are long-timescale tailing distortions, which can cause the effect of the current waveform to be influenced by historical waveforms applied before it, impairing the reproducibility of logic gates. Therefore, testing and compensating for Z-control line distortion is crucial for achieving high-fidelity quantum operations.

[0024] The existing technology, reference CN121212385A (a distortion correction optimization method and device for a quantum measurement and control link), measures the unit step response of the quantum measurement link in the time domain and then uses a deconvolution method to pre-compensate the desired waveform for distortion. This technology has the following problems: The higher the desired distortion accuracy of the quantum measurement and control link, the longer the entire experiment takes, and the testing time is directly proportional to the circuit response time. If multiple iterations are performed based on the residual response to improve the correction accuracy, the experimental time will increase exponentially, making it difficult to scale the time-domain test of a single step response to large-scale qubits. Furthermore, when studying the circuit response in conjunction with short-time distortion, the applicant often uses the sum of multiple e-exponents to fit the response of the quantum measurement and control system, but this fitting result deviates significantly from the data collected by the experimental system. In addition, deconvolution methods, including Fourier transform and inverse Fourier transform, are slow in calculating the predistorted waveform, thus increasing the compilation time during quantum circuit operation. Moreover, deconvolution methods require the entire quantum circuit to be determined in advance, which is incompatible with the operation mode of feedback-based quantum measurement and control links.

[0025] Based on the above issues, such as Figure 1 As shown, this application proposes a distortion correction optimization method for a quantum telemetry and control link (Z-link) to correct the distortion of the link, thereby ensuring the distortion-free transmission of signals and the fidelity and accuracy of operations in quantum computing and quantum simulation. The quantum telemetry and control link to be corrected is connected to a target qubit. Specifically, the distortion correction method of this application includes the following steps.

[0026] Step S10: Obtain a seed library containing multiple seeds; wherein each seed represents the pole model parameters of a digital filter used to perform distortion correction on a calibrated quantum measurement and control link, and the digital filter is determined based on a system function that incorporates the signal response characteristics of the quantum measurement and control link.

[0027] Among them, each sub-element represents the pole model parameters of the digital filter used to perform distortion correction on a calibrated quantum telemetry and control link. The digital filter is determined based on a system function that incorporates the signal response characteristics of the quantum telemetry and control link. This seed library is a continuously accumulating and expanding database that brings together historical optimization results from different quantum computing platforms, chips, and qubits.

[0028] As we've worked on more and more online platforms, we've accumulated distortion data for digital filters across numerous systems. The pole model parameters of different systems exhibit similarities and consistency. For example... Figure 2 A schematic diagram of the unit step response of multiple telemetry and control Z-links, where gray and green represent telemetry and control Z-links of different quantum computers.

[0029] Correspondingly, such as Figure 3 As shown in the figure, this embodiment provides an example diagram of a seed bank; Figure 2Each row in the table represents a seed characterization of the pole model parameters used for distortion correction.

[0030] Step S20: Execute the optimization algorithm based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, and apply a frequency control signal on the Z link of the target qubit according to the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit.

[0031] Specifically, for a Z-link to be tested, one or more seeds are selected from the seed library, and their pole model parameters are directly used as the initial values ​​of the optimization algorithm (such as the NM algorithm) or part of the initial population (such as the DE algorithm) to start the optimization process. During the iteration process, the optimization algorithm continuously adjusts the pole model parameters and uses the adjusted parameters to generate a predistortion waveform. The quality of the parameters is evaluated by measuring the quantum state fidelity of the target qubit.

[0032] Obtaining the fidelity of the quantum state of a target qubit requires testing. For example, a frequency control signal is applied to the Z-link of the target qubit, and logic gates for quantum state manipulation are applied to the XY-links of the target qubit. The fidelity of the quantum state of the target qubit is then tested. The logic gates are combinations of single quantum logic gates, exemplified by a randomized benchmark sequence.

[0033] It should be noted that the random benchmark sequence is a combination of quantum logic gates used in random benchmarking. Random benchmarking (RB) is a protocol used to evaluate the "average performance" of gate operations on quantum chips, aiming to determine the error probability of each gate operation in a computing environment. RB protocol quantum circuits can run on a real quantum chip or be simulated using a quantum simulator. When using a quantum simulator, ideally, the final result after executing the quantum circuit should be exactly the same as the initial state, thus increasing the accuracy of the test. Figure 4 The diagram shown illustrates the experimental principle of performing randomized benchmark tests. Figure 5 The diagram illustrates the results of random benchmarking under different conditions. It should be noted that random benchmarking refers to: preparing the qubit in the |0> state, then randomly selecting m group elements C1 from the Clifford group and applying them sequentially to the qubit; any group element C1 can also be called a Clifford gate, which can be obtained by combining basic physical gates, namely the X gate, X / 2 gate, Y gate, and Y / 2 gate; according to the definition of a group, there must exist an inverse in the Clifford group such that this series of operations is equivalent to a unit gate, which is then applied to the bit. If the operations are perfect, the bit should be in the |0> state; the equivalent unit gate is denoted as Cr.

[0034] The probability of a bit being in the |0> state is used as the fidelity of the RB sequence; for each m, the experiment is repeated k times to obtain the average value P_m of the sequence fidelity.

[0035] like Figure 4 As shown, a square wave is applied as a frequency control signal on the Z-link of the qubit, and a random reference sequence is applied on the XY-link. If the Z-link is not compensated, the tail of the square wave (as shown by the red dashed line in the figure) will cause the bit frequency to deviate from the operating point of the single-qubit logic gate, thus affecting the fidelity of the single-qubit logic gate. The RB test results are as follows: Figure 5 As shown in the green section. When using a pre-distorted waveform to accurately compensate for the square wave tail, the square wave will not affect the subsequent RB results, as shown in the image. Figure 5 As shown in orange. Figure 5 The blue area represents the RB result measured when no waveform is applied to the Z-link of the qubit.

[0036] As one implementation of the above embodiments, the spacing between the frequency control signal and the logic gate signal is set so that the impulse response of finite length has no effect on the quantum state change of the target qubit.

[0037] Without this interval, the compensation component used for distortion correction might persist into the XY pulse application period, interfering with the evolution of the quantum state through Z-XY coupling (such as residual ZZ coupling or stray coupling), leading to inaccurate measurement results. By ensuring that the distortion correction response is fully completed at the start of the XY operation, this scheme allows the fidelity measurement to reflect only the compensation effect of the Z-link distortion, improving the convergence accuracy of the entire optimization process and the effectiveness of the final correction parameters; simplifying the signal response types in the link also helps improve the fidelity of the quantum state of the qubit.

[0038] Step S30: Determine the pole model parameters when the fidelity meets the preset threshold as the target calibration parameters of the quantum measurement and control link to be calibrated.

[0039] When the optimization algorithm converges and the measured quantum state fidelity reaches a preset threshold (such as 99.9%), the corresponding pole model parameters are the optimal correction parameters for the Z-link under test.

[0040] This application utilizes a seed library to reuse the pole model parameters of digital filters used in previously distortion-corrected Z-links on new, similar links. This avoids starting from scratch to collect and fit the time-domain response, greatly simplifying the testing process and saving significant testing time. Furthermore, since the seed parameters are derived from real optimization data, their large sample size and higher reliability provide the optimization algorithm with a starting point closer to the global optimum. As the seed library gradually expands, the initial population and the actual parameters of the Z-link under test are more likely to overlap, significantly increasing optimization efficiency and convergence speed.

[0041] As one implementation of the above embodiments, the system function includes pure zero-point terms, real pole terms, and complex pole terms. The pole model parameters of each type include several real pole term parameters and complex pole term parameters. Among them, the pure zero-point terms represent the finite-length impulse response of the quantum measurement and control link, the real pole terms represent the exponential response of the quantum measurement and control link, and the complex pole terms represent the oscillation decay response of the quantum measurement and control link.

[0042] In this embodiment, the system response of the quantum measurement and control link is expressed as a function in the frequency domain as follows:

[0043] in, , and Each of these sub-items corresponds to a pure zero-point term, a real pole term, and a complex pole term, respectively, and their specific forms are determined by the signal response type. and These are the coefficients of the real pole terms and the coefficients of the complex pole terms.

[0044] In this embodiment, the system response of the quantum measurement and control link is expressed as a function in the time domain as follows:

[0045] The above function expression consists of polynomial terms, real pole terms, and complex pole terms, where the polynomial terms represent the impulse response of finite length ( The real pole term represents the exponential response ( The complex pole term represents the oscillatory damped response. ).

[0046] in, The coefficients of the polynomial terms are... The sampling interval (for example, it is 1 / 1.2ns on the AWG board). The proportionality coefficient of the real pole term. Let be the attenuation constant of the real pole term. For real poles, The proportionality constant of the complex pole term. For amplitude, Let be the attenuation constant of the complex pole term. Let be the oscillation period of the complex pole term. is the phase of the complex pole term; in addition, s, l, and k are constant values, and is the subscript value of ∑.

[0047] In the above function expression, the polynomial terms represent the FIR response, while the real and complex pole terms describe the IIR response. In this embodiment, the pole model parameters of each sub-sub include several real pole term parameters ( and ) and complex pole parameters ( , , , ).

[0048] It should be added that real pole parameters are several parameters of real poles, and complex pole parameters are several parameters of complex poles; each real pole includes 2 real pole parameters ( and Each complex pole includes four complex pole term parameters ( , , , The total number of pole model parameters for each seed depends on the number of real and complex poles included in that seed.

[0049] Example, Figure 3 In the example seed library, each row of data corresponds to 3 real poles (6 parameters in total) and 0 complex poles (0 parameters in total), so there are a total of 6 parameters. When the number of real or complex poles doubles, the number of parameters in the corresponding pole terms also doubles, and the number of pole model parameters in each seed also increases accordingly; for example, if each seed includes 3 real poles and 2 complex poles, then each seed should have 14 parameters.

[0050] The response of the corresponding digital filter (IIR filter) can be determined by the pole model parameters mentioned above. For different pole model parameters, the IIR filter has corresponding coefficients to achieve different degrees of distortion correction.

[0051] In addition, the number of real poles and complex poles is several. When the number of real poles and complex poles is 1, the pole model parameters of each sub-sub include the above 6 parameters. When the number of real poles or complex poles is doubled, the number of corresponding pole term parameters is doubled, and the pole model parameters in each sub-sub also increase accordingly.

[0052] This embodiment starts from the system transfer function and derives a parameterized model to describe the distortion behavior of the Z-link. The parameterized model accurately matches the signal response mechanism of the real physical link. Compared with a single function form (such as pure exponential), this hybrid model can fit various complex distortion behaviors with high accuracy with fewer parameters. This makes the parameters of the seed established based on this model have clear physical meaning, and the parameter similarity and consistency (such as similar time constants) between different links are higher, making seed reuse more effective.

[0053] As one implementation of the above embodiments, the optimization algorithm includes the Nelder-Mead algorithm or the Differential Evolution algorithm. When the NM algorithm is selected, a single initial value (i.e., the best seed selected from the seed bank) is provided. The algorithm constructs a simplex and searches in the parameter space based on operations such as reflection, expansion, and contraction to find the parameters that optimize the fidelity (or the cost function derived from the fidelity). When the DE algorithm is selected, an initial population needs to be provided. This invention uses NIND best seeds selected from the seed bank as the initial population; the DE algorithm iteratively evolves the population through mutation, crossover, and selection operations to find the global optimum.

[0054] The NM algorithm converges quickly and is suitable for unimodal problems or situations where high-quality initial values ​​are available; the DE algorithm has strong global search capabilities, is suitable for multimodal problems, and can avoid getting trapped in local optima. A seed library can provide high-quality initial parameter values ​​for both, allowing for flexibility in choosing between different algorithms.

[0055] As one implementation of the above embodiments, when the optimization algorithm includes a differential optimization algorithm, before executing the optimization algorithm based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, the method further includes the following steps.

[0056] Step S101: Determine the target population size for the differential evolution algorithm; Step S102: Based on the pole model parameters of all seeds in the seed library, perform a fidelity test of the quantum state of the qubit on each quantum measurement and control link; Step S103: Determine the number of seeds for the target population based on the size of the fidelity test results as the first seed bank.

[0057] As the number of seeds in the seed bank continues to grow, it's impossible to set the population size in the experimental options to match the number of seeds in the seed bank. Assume the seed bank has accumulated S = 150 pole parameter seeds, and the population size NIND for the DE algorithm is set to 20. A new preliminary experiment is added: for each bit, optimization experiments are run sequentially using the parameters from the seed bank, and the results are sorted from smallest to largest. Finally, NIND of the best seeds are selected from the seed bank of size S as the initial population for the subsequent formal optimization of that bit.

[0058] For example, suppose the total size of the seed bank is S (e.g., 150), and the initial population size required by the DE algorithm is NIND (e.g., 20). For the Z link to be tested, do not immediately begin formal optimization. Instead, conduct a rapid preliminary experiment: sequentially take each seed (S in total) from the seed bank, apply its pole model parameters directly to the link (i.e., as coefficients of the digital filter), and then perform a rapid fidelity test (e.g., perform a simple Ramsey interferometry experiment, or...). Figure 4 (Representing the same experiment); record the fidelity results for each seed and sort them from highest to lowest fidelity; from the sorted list, select the top NIND seeds (e.g., the top 20) with the highest fidelity and combine them to form a new, customized "first seed pool". Finally, use these NIND seeds as the initial population for the DE algorithm to begin formal parameter optimization iterations.

[0059] For example, the optimization algorithm is executed based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, including: the optimization algorithm is executed based on the pole model parameters of several seeds in the first seed library as the initial values ​​of the parameters to be optimized.

[0060] Using the aforementioned preliminary experiment, when the seed pool size S is much larger than the population size NIND required by the optimization algorithm, a lightweight preliminary experiment determines an optimal initial population for the specific test link, rather than randomly selecting or simply reusing a single seed. This ensures that the DE algorithm starts its search from a very good and diverse starting point, greatly improving optimization efficiency and convergence speed. Furthermore, this preliminary experiment is highly parallelizable, allowing simultaneous testing of all bits without significantly increasing the total testing time.

[0061] As one implementation of the above embodiments, applying a frequency control signal on the Z-link of the target qubit according to the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit includes the following steps.

[0062] Step S201: Apply a frequency control signal for loading a predistortion waveform to the Z-link of the target qubit according to the optimized pole model parameters; wherein the predistortion waveform is generated by a digital filter based on the pole model parameters.

[0063] The set of pole model parameters provided by the optimization algorithm is fed into a digital filter module, and the system generates an ideal target frequency control waveform (e.g., a square wave pulse for implementing Z-rotation). This ideal waveform serves as the input to the digital filter, which calculates the output based on the pole model parameters to obtain a predistorted waveform. This predistorted waveform is then loaded onto the arbitrary waveform generator (AWG) of the Z-link, ready to be applied to the Z-control line of the target qubit.

[0064] Step S202: Apply logic gate signals for quantum state operations to the XY control lines of the target qubit.

[0065] Under the coordination of the timing controller, the aforementioned predistortion waveform is first applied as a frequency control signal to the Z-link of the target qubit. The signal is used to change the frequency of the qubit to achieve specific phase accumulation. After a precise time interval following the application of the predistortion waveform, a logic gate signal is applied to the target qubit through the XY control line. This signal is used for the resonant microwave pulse for quantum state operation, such as a π pulse (X_gate) or a π / 2 pulse.

[0066] Step S203: Obtain the fidelity of the quantum state output by the target qubit.

[0067] After executing the above sequence of operations, the target qubit is read out. By repeating the sequence multiple times (e.g., 1024 or 4096 times), the probability of the qubit being in the target quantum state is calculated, and the measured value (e.g., the operation fidelity of X_gate) is used as the fitness value of the set of pole model parameters and returned to the optimization algorithm.

[0068] Through steps S201-S203, a closed-loop evaluation system with the final quantum operation fidelity as the direct optimization target is constructed. Compared with the method of indirectly evaluating the distortion correction effect by measuring electrical signals (such as observing the step response with an oscilloscope), this scheme directly measures the response of the qubit. Therefore, the optimized pole model parameters can maximize the fidelity of the quantum gate operation, avoid the deviation between indirect evaluation and actual performance, and ensure the true effectiveness of the correction effect.

[0069] As one implementation of the above embodiments, the digital filter generates a predistortion waveform based on pole model parameters, including the following steps.

[0070] Step S2011: Determine the coefficients of the digital filter based on the pole model parameters.

[0071] Step S2012: The digital filter generates a predistortion waveform based on the coefficients.

[0072] Specifically, in the field of digital signal processing, digital filters and system functions satisfy an inverse relationship, meaning the digital filter equals 1 / coefficient function. The coefficients of the denominator and numerator polynomials of the system function can be used to determine the forward and feedback filter coefficients of the digital filter, respectively. These coefficients are determined by the parameters of the pure zeros, real poles, and complex poles. In practical implementation, high-order digital filters may face risks to numerical accuracy; therefore, they can be transformed into a series of quadratic filters with a direct transpose structure.

[0073] By transforming pole model parameters into filtering coefficients of the physical waveform, the pole model parameters are no longer abstract mathematical quantities, but rather actionable control signals that can affect the qubits. Standardized coefficient transformation algorithms ensure comparability and reproducibility between different parameters. Simultaneously, the generation of predistorted waveforms is performed in real-time by the digital filter, allowing for rapid evaluation of each new set of parameters during optimization without recompiling or downloading complex waveform files, thus significantly accelerating the optimization iteration process. When implemented as FPGA hardware, the digital filter can perform real-time correction of waveform distortion in the quantum telemetry and control link.

[0074] In one implementation, the method further includes updating the seed library by using the target calibration parameters as new seeds. After obtaining the target calibration parameters for a Z-link to be tested through step S30, the system automatically adds the set of parameters back to the seed library as new seeds. The seed library can be designed as a dynamically growing database, with each record containing a complete set of pole model parameters and optional metadata, such as source platform, chip ID, optimization time, final fidelity, etc.

[0075] After each measurement and control Z-link optimization is completed, the obtained seed is retained and used for subsequent optimization tasks, realizing a self-reinforcing loop of the seed library. As the number of online platforms and optimized links increases, the size S of the seed library continues to grow, and its parameter space becomes more complete. This makes it easier for subsequent links to be tested to match highly similar initial parameters from the seed library, thereby further increasing optimization efficiency and accelerating convergence speed.

[0076] like Figure 6 As shown, based on the same application concept, this embodiment also provides a distortion correction and optimization device for a quantum measurement and control link. The quantum measurement and control link to be corrected is connected to a target qubit and includes the following modules.

[0077] The seed acquisition module is used to acquire a seed library containing multiple seeds; wherein, each seed represents the pole model parameters of the digital filter used to perform distortion correction on a calibrated quantum measurement and control link, and the digital filter is determined based on a system function that integrates the signal response characteristics of the quantum measurement and control link; The algorithm execution module is used to execute the optimization algorithm based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, and to apply a frequency control signal on the Z link of the target qubit according to the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit. The parameter determination module is used to determine the pole model parameters when the fidelity meets the preset threshold as the target correction parameters of the quantum measurement and control link to be corrected.

[0078] The seed acquisition module and parameter determination module mentioned above can be functional modules integrated into a classical computer, while the algorithm execution module can be a specific measurement device, such as a quantum computing measurement and control system, used to output various control signals and perform measurements.

[0079] The device is a physical entity or software system that implements the above method. The various modules work together to automatically complete the entire process from seed library acquisition and optimization to final parameter determination, achieving a high degree of automation in distortion correction.

[0080] Based on the same concept, this embodiment also provides a quantum computer that uses the distortion correction optimization method as described in any of the above embodiments to correct the quantum measurement and control link connecting the qubits, or includes the distortion correction optimization device as described in the above embodiments.

[0081] This quantum computer integrates the seed library and optimization process of this invention, giving it significant advantages in multi-qubit, cross-platform calibration tasks. Specifically, the calibration time is greatly reduced when a new chip is deployed; the system achieves higher stability after long-term operation because it can be periodically recalibrated using the seed library; and the fidelity and consistency of the overall quantum operations are effectively guaranteed.

[0082] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, which can take the form of a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email sending and receiving device, game console, tablet computer, wearable device, or any combination of these devices.

[0083] In a typical configuration, a computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0084] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0085] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0086] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0087] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0088] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to limit the one or more embodiments of this specification. The singular forms “a,” “described,” and “the” used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more associated listed items. It should be understood that although the terms first, second, third, etc., may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, first information may also be referred to as second information without departing from the scope of one or more embodiments of this specification, and similarly, second information may also be referred to as first information. Depending on the context, the word “if” as used herein may be interpreted as “when,” “when,” or “in response to a determination.” The above descriptions are merely preferred embodiments of one or more embodiments of this specification and are not intended to limit the one or more embodiments of this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments of this specification should be included within the scope of protection of one or more embodiments of this specification.

[0089] The above description, based on the embodiments shown in the drawings, details the structure, features, and effects of this application. The above description is only a preferred embodiment of this application, but this application does not limit the scope of implementation to what is shown in the drawings. Any changes made in accordance with the concept of this application, or modifications to equivalent embodiments, that do not exceed the spirit covered by the specification and drawings, should be within the protection scope of this application.

Claims

1. A distortion correction optimization method for a quantum measurement and control link, wherein the quantum measurement and control link to be corrected is connected to a target qubit, characterized in that, The method includes: A seed library containing multiple seeds is obtained; wherein each seed represents the pole model parameters of a digital filter used to perform distortion correction on a calibrated quantum measurement and control link, the digital filter being determined based on a system function that incorporates the signal response characteristics of the quantum measurement and control link; The optimization algorithm is executed based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, and the frequency control signal is applied on the Z link of the target qubit according to the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit. The pole model parameters that satisfy the preset threshold are determined as the target calibration parameters of the quantum measurement and control link to be calibrated.

2. The distortion correction optimization method for quantum measurement and control links according to claim 1, characterized in that, The system function includes pure zero-point terms, real pole terms, and complex pole terms. The pole model parameters of each seed include several real pole term parameters and complex pole term parameters. The pure zero-point terms represent the finite-length impulse response of the quantum measurement and control link, the real pole terms represent the exponential response of the quantum measurement and control link, and the complex pole terms represent the oscillation decay response of the quantum measurement and control link.

3. The distortion correction optimization method for the quantum telemetry and control link according to claim 1, characterized in that, The optimization algorithm includes the Nelder-Mead algorithm or the differential evolution algorithm.

4. The distortion correction optimization method for the quantum measurement and control link according to claim 3, characterized in that, When the optimization algorithm includes a differential optimization algorithm, before executing the optimization algorithm based on the pole model parameters of several seeds as initial values ​​of the parameters to be optimized, the method further includes: Determine the target population size for the differential evolution algorithm; Based on the pole model parameters of all seeds in the seed library, the fidelity test of the quantum state of the qubit is performed on each quantum measurement and control link; The first seed bank is determined by sorting the seeds according to the size of the fidelity test results to determine the target population size mentioned above. The optimization algorithm is executed using the pole model parameters based on several seeds as initial values ​​for the parameters to be optimized, including: The optimization algorithm is executed based on the pole model parameters of several seeds in the first seed library as the initial values ​​of the parameters to be optimized.

5. The distortion correction optimization method for quantum measurement and control links according to claim 1, characterized in that, The step of applying a frequency control signal to the Z-link of the target qubit based on the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit includes: A frequency control signal for loading a predistortion waveform is applied to the Z-link of the target qubit according to the optimized pole model parameters; wherein the predistortion waveform is generated by the digital filter based on the pole model parameters; A logic gate signal for quantum state operation is applied to the XY control line of the target qubit; Obtain the fidelity of the quantum state output by the target qubit.

6. The distortion correction optimization method for quantum measurement and control links according to claim 5, characterized in that, The digital filter generates the predistortion waveform based on the pole model parameters, including: The coefficients of the digital filter are determined based on the pole model parameters. The digital filter generates a predistorted waveform based on the coefficients.

7. The distortion correction optimization method for a quantum telemetry and control link according to claim 6, characterized in that, The spacing between the frequency control signal and the logic gate signal is set so that the impulse response of finite length has no effect on the quantum state change of the target qubit.

8. The distortion correction optimization method for quantum measurement and control links according to claim 1, characterized in that, The method further includes: The target correction parameters are used as the new seed to update the seed library.

9. A distortion correction and optimization device for a quantum measurement and control link, wherein the quantum measurement and control link to be corrected is connected to a target qubit, characterized in that, The device includes: A seed acquisition module is used to acquire a seed library containing multiple seeds; wherein each seed represents the pole model parameters of a digital filter used to perform distortion correction on a calibrated quantum measurement and control link, and the digital filter is determined based on a system function that integrates the signal response characteristics of the quantum measurement and control link; The algorithm execution module is used to execute the optimization algorithm based on the pole model parameters of several seeds as the initial values ​​of the parameters to be optimized, and to apply a frequency control signal on the Z link of the target qubit according to the optimized pole model parameters to obtain the fidelity of the quantum state of the target qubit. The parameter determination module is used to determine the pole model parameters when the fidelity meets a preset threshold as the target correction parameters of the quantum measurement and control link to be corrected.

10. A quantum computer, characterized in that, The distortion correction optimization method for the quantum measurement and control link as described in any one of claims 1-8 is used to correct the quantum measurement and control link connecting the qubits, or the method includes the distortion correction optimization device for the quantum measurement and control link as described in claim 9.

Images