Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This is the initial Office Action based on the application filed March 26, 2025. Claims 1-12 and 15 are presented for examination and have been considered below.
Specification
The disclosure is objected to because of the following informalities:
Page 5, Line 9, “functions or objective variables.to optimize” should be “functions or objective variables to optimize”.
Page 9, Line 25, “10-“ should be “10.” to match with the consistency of the formatted list.
Appropriate correction is required.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1 and 5 are rejected under 35 U.S.C. 103 as being unpatentable over Elfving (US 2025/0299084) in view of Matsuura (WO 2019/241879)
As per claim 1, Elfving teaches a method for quantum algorithm generation or orchestration for solving optimization problems (Abstract, “methods and systems determine a solution for an optimization problem”) comprising the steps of: defining an optimization problem as a set of parameters and constraints (paragraph 18, "The method comprises receiving or determining a description of the optimisation problem", "The determination may comprise execution of gate operations defined by the first parametric quantum circuit using the optimised first parameters or a derivative thereof and acquisition of measurement data associated with an output state of the quantum computer system"), a machine where the problem should be executed (paragraph 39, "This optimization problem may be solved by a non-classical computer, such as a quantum computer”), and adapting the optimization problem data to the quantum algorithm (paragraph 25, "(QEL) algorithm learns model based on the available data"). Elfving does not disclose the determining of an optimum set of quantum/traditional algorithms, comprising at least one quantum algorithm, for the problem defined and its hyperparameters by determining a strategy for solving selecting from: annealing based, gate based, or a black box based approach, and the providing of the adapted optimization problem to the quantum algorithm for being solved by using reward functions or accuracy variables for optimization.
However, Matsuura discloses in a computer-implemented method for solving an optimization problem, the operation for solving utilizes an optimizer selected from a group of methods, which include Bayesian, gradient-free, and a black-box optimization method (paragraph 8). A reinforcement learning procedure within the implemented method involves an agent that provides a reward function based on the procedure’s interaction with an environment. The agent’s goal is to enhance or maximize the reward function as the method continues to be optimized (paragraph 72). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to implement an optimal method type to solve an optimization problem based on the parameters and also a reward system in the optimization algorithm as this helps users find an optimal state of the given problem (Matsuura, paragraph 37) in a significant reduced amount of computational time (Matsuura, paragraph 6).
As per claim 5, Elfving, in view of Matsuura, teaches the method of Claim 1. Elfving further teaches the step of determining an optimum set of quantum and/or traditional algorithms involves the use of: quantum hardware (paragraph 377, "quantum processors"), traditional hardware platforms (CPUs and GPUs) (paragraph 54, "computer program instructions may be provided to a processor, in particular a microprocessor or central processing unit (CPU), or graphics processing unit (GPU)") and/or photonic or neuromorphic computing (paragraph 381, "the one or more quantum processors may comprise a set of continuous variable systems, such as optical or photonic quantum computers").
Claims 2-4, 6-8 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over the combination of Elfving (US 2025/0299084) and Matsuura (WO 2019/241879A1), in view of Sharma (US 2025/0209368).
As per claim 2, Elfving, in view of Matsuura, does not disclose a step of automatically generating an API for providing user access management, security protocols and monetization mechanisms.
However, Sharma discloses in a QAOA system the creation of OpenQAOA, which provides an API to build and run quantum algorithms and workflows on multiple backends (Fig. 1; paragraph 24). Within the API is a resource estimator and profile module (Fig. 1, item 155) which provides a user with information, including cost of execution and financial estimates, to assist users in evaluating resources required to meet their objections (paragraph 108). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to incorporate an API for user management in a quantum correction system as this improves scalability and efficiency of optimization and the creation of optimization workflows allow the user to find the most optimal and efficient quantum processing path for optimization algorithms (Sharma, paragraph 28).
As per claim 3, Elfving, in view of Matsuura, does not disclose a step of generating a synthetic data Set Generation for algorithm testing.
However, Sharma discloses that the OpenQAOA system can provide fast and diverse methods to classically simulate quantum approximate optimization algorithms (QAOA) under different conditions, such as noisy and noiseless computations (paragraph 30). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to have the optimization algorithm generate synthetic data as this advantageously enables assessment of the simulated QAOA’s performance and acquire useful information about any noise characteristics of the quantum device when the algorithm is executed on real quantum computers (Sharma, paragraph 30).
As per claim 4, Elfving, in view of Matsuura, does not disclose a step of the method using hyperparameters for algorithm testing.
However, Sharma discloses within the OpenQAOA core is a set of pre-defined settings (Fig. 1, item 130), which contains specifiable default values and selection of many hyperparameters for quantum approximate optimization algorithm QAOA workflows (paragraph 84). These settings let the user provide the associated hyperparameters for the computation to be fully executed (paragraph 82). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to have hyperparameters in the quantum optimization method as they reduce time consumption of selecting specific parameters and are already verified to be optimal compared to a random selection of parameters (Sharma, paragraph 84).
As per claim 6, Elfving, in view of Matsuura, does not disclose step of determining optimum algorithms, machines, and hyperparameters for a defined problem by using artificial intelligence with past data or the synthetic data Set Generation.
However, Sharma discloses that in the OpenQAOA system an AutoVQA (Fig. 6B, item 156) is used to create optimized VQAs can perform analytics to determine the best choices of parameterizations, optimizer, and the heuristics to improve the quality of the optimization result (paragraph 169). The AutoVQA uses a method where it compares problem instances with previous problem instances through machine learning (paragraph 114). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to use artificial intelligence trained with past data to find the optimal conditions for a given problem as they improve the usability and performances of building quantum optimization algorithms (Sharma, paragraph 15).
As per claim 7, Elfving, in view of Matsuura, teaches in the method a process of a continuous updating system for a defined problem (paragraph 94, “iterative process continues until classical optimizer converges or solution of acceptable quality is found”) using artificial intelligence trained with new incoming data (paragraph 96, “recommendation system created trained model based on training data then suggests new data that have the desired features”), but does not disclose that the system is used for updating the optimum algorithms, machines, and hyperparameters.
However, Sharma discloses within the OpenQAOA system a classical quantum loop (Fig. 6C, item 630) that creates the variational quantum algorithms based on a given agnostic representation (Fig. 6B, item 622). The classical quantum loop contains a circuit builder (Fig. 6C, item 632) which sends the representation to a backend device (Fig. 6C, item 636) to generate a cost function (Fig. 6C, item 637). The classical optimizer (Fig. 6C, item 638) analyzes the cost function and creates new parameters, which are used to update the circuit in the circuit builder. The updated circuit then gets sent to the backend device for reevaluation and the loop continues until the optimization process is complete (paragraph 155). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to use artificial intelligence trained with new data to find the optimal conditions for a given problem as they improve the usability and performances of building quantum optimization algorithms (Sharma, paragraph 15).
As per claim 8, Elfving, in view of Matsuura teaches in the method a process of a continuous updating system for a defined problem (paragraph 94, “iterative process continues until classical optimizer converges or solution of acceptable quality is found”), but does not disclose the system being used to update the optimum algorithms, machines, and hyperparameters by adding new algorithms to an existing pool of available algorithms for solving.
However, Sharma discloses in the OpenQAOA core there is an input model (Fig. 1, item 115) that includes a user-defined problem and a pre-defined VQA state (paragraph 14), which gets sent to the enabler API (Fig. 1, item 140), then to the AutoVQA (Fig. 1, item 156). Between the enabler API and AutoVQA is a classical quantum loop (Fig. 6C, item 630) that creates the variational quantum algorithms based on a given agnostic representation (Fig. 6B, item 622). The classical quantum loop contains a circuit builder (Fig. 6C, item 632) which sends the representation to a backend device (Fig. 6C, item 636) to generate a cost function (Fig. 6C, item 637). The classical optimizer (Fig. 6C, item 638) analyzes the cost function and creates new parameters, which are used to update the circuit in the circuit builder. The updated circuit then gets sent to the backend device for reevaluation and the loop continues until the optimization process is complete (paragraph 155). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to add new algorithms to an existing pool of algorithms to find the optimal conditions for a given problem as they improve the usability and performances of building quantum optimization algorithms (Sharma, paragraph 15).
As per claim 15, Elfving, in view of Matsuura, does not disclose a step of the method using hyperparameters for algorithm testing.
However, Sharma discloses within the OpenQAOA core is a set of pre-defined settings (Fig. 1, item 130), which contains specifiable default values and selection of many hyperparameters for QAOA workflows (paragraph 84). These settings let the user provide the associated hyperparameters for the computation to be fully executed (paragraph 82). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to have hyperparameters in the quantum optimization method as they reduce time consumption of selecting specific parameters and are already verified to be optimal compared to a random selection of parameters (Sharma, paragraph 84).
Claims 9 and 10 are rejected under 35 U.S.C. 103 as being unpatentable over the combination of Elfving (US 2025/0299084) and Matsuura (WO 2019/241879A1), in view Davis (US 2023/0020389).
As per claim 9, Elfving, in view of Matsuura, does not disclose a step of solving at least one subproblem defined to at least one auxiliary machine, enabling parallel computing.
However, Davis discloses a quantum logic circuit that generates execution tasks or subproblems (paragraph 86) after it has been scattered. These execution tasks can be stored in a scheduler module (Fig. 5, item 514) that is located within a cluster server (Fig. 10; paragraph 94), where the scheduler module produces an execution schedule (paragraph 96). An execution schedule includes a set of execution tasks (paragraph 115) and in some instances there can be a second execution schedule (paragraph 115), where a computing system can run both first and second execution schedules in parallel (paragraph 115). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to perform parallel processing for subproblems as this improves or optimizes computing performance of the cluster server’s operation (Davis, paragraph 37).
As per claim 10, Elfving, in view of Matsuura, does not disclose a step where at least one subproblem is performed by asynchronous management.
However, Davis discloses a quantum logic circuit that generates execution tasks or subproblems (paragraph 86) after it has been scattered. The execution of the execution tasks on shared hardware may be executed in an asynchronous manner (paragraph 115). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to perform asynchronous operations for subproblems as this improves or optimizes computing performance of the cluster server’s operation (Davis, paragraph 37).
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over the combination of Elfving (US 2025/0299084) and Matsuura (WO 2019/241879A1), in view of Macready (US 2011/0047201).
As per claim 11, Elfving, in view of Matsuura, does not disclose hyperparameters of the annealing solver when selected, which includes annealing time (selecting the time for the quantum annealer to evolve an initial Hamiltonian to the problem Hamiltonian), annealing schedule (selecting how the Hamiltonian evolves over time), and the chain strength of physical qubits (selecting the strength of the connections between the qubits).
However, Macready discloses within an apparatus that solves a discrete optimization problem using a quantum processing system, the system has an annealing process which has an associated annealing time, which can be selected for the quantum state to find its lowest energy configuration (paragraph 138) for the Hamiltonian to evolve from an initial to a problem state (paragraph 137). The system’s memory module (Fig. 10, item 1020) has a user interface module (Fig. 10, item 1022), that can allow a user to define an optimization problem through adjusting run-time control parameters, such as an evolution schedule (paragraph 159). The system also contains quantum devices (Fig. 9A, 9B) which represent the qubits that are coupled together by a coupling strength (paragraph 127), for which the coupling strength represents the coefficients linking paired variables. The Hamiltonian can be affected by changing the strength of the couplings that link the quantum devices together (paragraph 137). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to include the hyperparameters of the annealing solver when selected as they are well suited and effective for discrete and combinatorial optimization problems compared to other methods (Elfving, paragraphs 82-83).
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over the combination of Elfving (US 2025/0299084) and Matsuura (WO 2019/241879A1), in view of Chamberland et. al. (herein Chamberland, US 12007835).
As per claim 12, Elfving, in view of Matsuura, does not disclose the hyperparameters of a gate based solver when selected, which includes circuit depth (selecting a number of gate layers chosen), gate selection (selecting types of gates used and their arrangement), error correction code (selecting quantum error correction code, along with its code distance and code rate), and initial state preparation (selecting the method used to prepare initial state of qubits).
However, Chamberland discloses a quantum algorithm that uses Pauli-based computations to reduce lattice surgery measurements. The quantum device associated with the Pauli-based computation can resemble a grid of qubits that can be initialized in various ways to form the ideal configuration of the computer for making these computations (Col. 7, lines 55-60). The Pauli-based quantum computation can execute, with a given quantum algorithm, a number of T-gates with a number of depths (Col. 9, lines 16-21). The computation can also perform multi-qubit non-Clifford gates within Pauli measurements (which can be performed based on the arrangement of multi-qubit Pauli operators) (Col. 6, lines 25-31). Pauli operators may be encoded into codewords of given classical error-correcting code (abstract), for which the codewords can represent a classical [n, k, d] code, for which the d represents the code distance parameter (Col. 7, lines 5-10), and the effectiveness can be determined based of the code rate ratio k bits of useful information / n bits of data (Col. 25, lines 9-18). This would have been obvious for a person having ordinary skill in the art at the time the invention was effectively filed to include the hyperparameters of the gate solver when selected as they are advantageous for reducing space-time costs for fault-tolerant quantum error correction protocols (Chamberland, Col. 7, lines 20-24).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Amin (US 2011/0060710) discloses methods of using a hybrid classical and quantum system to solve computational problems. The hybrid system uses a Hamiltonian to perform annealing or adiabatic computations and processors to refine the first solution from the Hamiltonian and provide an optimal solution.
Hastings (US 2017/0330101) discloses a machine learning method to train a quantum algorithm to find the highest overlap with the optimal state of a given optimization problem, rather than looking for a quantum state which solves the problem.
McMahon (US 11526795) discloses a variational quantum algorithm that is processed through different types of quantum processing units, which are run in an iterative manner to optimize the circuit parameters for their respective circuits to find an ideal optimization solution.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KHANG HUU NGUYEN whose telephone number is (571)270-5978. The examiner can normally be reached Monday-Friday (8:00am - 5:00pm) ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Mark Featherstone can be reached at 571-260-3750. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/KHANG HUU NGUYEN/Examiner, Art Unit 2111
/MARK D FEATHERSTONE/Supervisory Patent Examiner, Art Unit 2111
Prosecution Insights