DETAILED ACTION
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 .
Claims 1-20 are pending for examination.
Priority
Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 08/13/2024 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-4 and 6-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Huijgen (Training Quantum Boltzmann Machines with the β-Variational Quantum Eigensolver, 17 Apr 2023).
Regarding claim 1, Huijgen teaches a method for performing machine learning using quantum computing hardware, the method comprising:
providing a model comprising a Quantum Boltzmann machine (QBM) with a Hamiltonian ansatz having a set of operators and a set of parameters (see respective Quantum Boltzmann machine in abstract: "The quantum Boltzmann machine (QBM) is a generative machine learning model for both classical data and quantum states. [ ... ] For the cases considered here, the obtained QBMs can model the target to high fidelity.", see β-VQE ansatz, which is a form of Hamiltonian ansatz in Figure 1, legend: "In the inner loop we optimize the variational free energy F of the β-VQE ansatz ρθ,φ with the QBM Hamiltonian Hw.");
performing a first stage of training the model against data from a target using a selected subset of the set of operators to obtain optimized values for a subset of the set of parameters, wherein the first stage of training is performed on classical binary computing hardware to provide a partly trained model (see the inner loop (i.e. first stage of training) of the nested-loop algorithm, wherein the β-VQE ansatz is optimized to approximate the Quantum Boltzmann Machine state involving classical network optimization (e.g. the classical part of the β-VQE ansatz, which defines the probability distribution optimized using a classical algorithms i.e. gradient descent) and parameter tuning (e.g. the parameters of both the classical network and the quantum circuit are optimized in this step. Notably, while quantum circuits are involved, the optimization of the ansatz itself is driven by classical optimization) in page 3, col. 2, paragraph 3: "The algorithm starts with a simple ansatz for the QBM Hamiltonian Hw, e.g., a Heisenberg XXZ model or a random spinglass model. In the inner loop, the β-VQE ansatz ρθ,φ is trained to represent the QBM σw by minimizing Eq. (6)", and equation 6); and
performing a second stage of training the model against data from the target using a larger subset of the set of operators to obtain optimized values for a larger subset of the set of parameters for the model, wherein the second stage of training is performed using quantum computer hardware, (see the outer loop (i.e. second stage of training), wherein the nested-loop algorithm trains the QBM to minimize the quantum relative entropy between the density matrix of the QBM and the target data. This stage requires calculating quantum properties, such as expectation values of the QBM Hamiltonian, involving quantum state evaluations (e.g. the quantum computer evaluates expectation values of the Hamiltonian -using the current QBM parameters- based on the trained β-VQE ansatz) and QBM model optimization (e.g. the QBM's parameters (weights of the Hamiltonian operators) are adjusted iteratively using the approximated statistics from the β-VQE) in page 3, col. 2, paragraph 3: "In the outer loop, ρθ,φ is used to compute approximate QBM statistics tr(Hr σw) ≈ tr(Hr ρθ*,φ*). These are used for training the QBM by gradient descent on Eq. (3).", and equation 3, wherein the outer loop involves the full QBM training and works with more data than the inner loop, which only operates on the truncated β-VQE approximation, i.e. the outer loop (second stage) refines the QBM by using more data, while the inner loop (first stage) optimizes the β-VQE on a smaller data subset, see page 3, col. 2, paragraph 2: "In the following, we refer to this as truncatedrank β-VQE. As an immediate consequence, the gradients for the variational free energy Eqs. (7) and (8) are also truncated for this model. This means that we can heuristically choose a small R so that the optimization can be performed at a reduced computational cost.", and equation 10)
and wherein the optimized values from the first stage of training are used to initialize corresponding parameters for the second stage of training (see ρθ,φ (optimized parameter values) from the first stage (inner loop) of training are used to initialise the corresponding parameters for the second stage (outer loop) of training, in page 3, col. 2, paragraph 3: "The algorithm starts with a simple ansatz for the QBM Hamiltonian Hw, e.g., a Heisenberg XXZ model or a random spinglass model. In the inner loop, the β-VQE ansatz ρθ,φ is trained to represent the QBM aw by minimizing Eq. (6). In the outer loop, ρθ,φ is used to compute approximate QBM statistics tr(Hr σw) ≈ tr(Hr ρθ*,φ*). These are used for training the QBM by gradient descent on Eq. (3).").
Regarding claim 2, Huijgen teaches the method of claim 1, including iterating the second stage of training with a larger subset of operators and/or a larger subset of parameters in each iteration, to provide a trained Quantum Boltzmann machine in which a difference in expectation values between a target and the model is iteratively reduced (see QBM iterations in Figure 1).
Regarding claim 3, Huijgen teaches the method of claim 1, wherein the first stage of training trains the model using quantum relative entropy between the model and the target (see outer loop (second stage) minimizing the quantum relative entropy between the QBM and the target density matrix, wherein the inner loop (first stage) provides an approximation for the QBM through β-VQE, which is used to calculate the expectation values needed for QBM updates in abstract, lines 2-3).
Regarding claim 4, Huijgen teaches the method of claim 3, wherein gradients of the quantum relative entropy are determined with respect to expectation values for the model and the target (see equation 4, and page 2, col. 2, last paragraph).
Regarding claim 6, Huijgen teaches the method of claim 1, wherein parameters which are not in the selected subset of the operators are maintained at zero during the first stage of training (see truncated-rank β-VQE in page 3, col. 2, paragraph 2).
Regarding claim 7, Huijgen teaches the method of claim 1, wherein the selected subsets of operators and parameters use substantially all computational resources from the classical binary computer hardware (see abstract, method as heuristic training to overcome computationally intractable QBM expectation values for large models, which implies substantial use of all computational resources from the classical computer hardware).
Regarding claim 8, Huijgen teaches the method of claim 1, further comprising extending the Hamiltonian ansatz from the subsets of operators and parameters of the first stage to the subsets of the operators and parameters of the second stage (see warm-up strategy in page 2, col. 1, paragraph 4, and see nested loop
wherein the QBM ansatz evolves from a simplified model in the inner loop to a more detailed model in the outer loop section D, and page 5, col. 2, first paragraph).
Regarding claim 9, Huijgen teaches the method of claim 1, wherein the second stage of training is performed on the quantum computing hardware with respect to all the parameters (see full QBM state in Figure 3).
Regarding claim 10, Huijgen teaches the method of claim 1, wherein the second stage of training includes optimizing quantum relative entropy with respect to all the parameters by computing Gibbs expectation values on the quantum computing hardware (see computing Gibbs expectation values in page 6, col. 2, paragraph 2).
Regarding claim 11, Huijgen teaches the method of claim 10, wherein the first stage of training comprises training the model using quantum relative entropy between the model and the target to provide the partly trained model comprising a Quantum Boltzmann Machine, and wherein the second stage of training comprises sampling the partly trained model by preparation of Gibbs states and the computing of Gibbs expectation values, wherein each sampling of the model comprises preparation of a Gibbs state and computation of Gibbs expectation values on the quantum computing hardware (see approximating the Gibbs state expectation values of the QBM using the β-VQE ansatz (i.e. partly trained model), wherein each sampling of the model (in the outer loop) comprises preparing an approximate Gibbs state (via the β-VQE ansatz) and calculating Gibbs expectation values, in page 6, col. 2, paragraph 2).
Regarding claim 12, Huijgen teaches the method of claim 1, wherein the second stage of training includes performing a stochastic gradient descent (see gradient descent for optimization using sampling, i.e., a stochastic gradient descent in page 2, col. 1, last paragraph, and page 3, col. 2, paragraph 2).
Regarding claim 13, Huijgen teaches the method of claim 1, wherein the second stage of training involves T iterations each involving N samples, wherein N×T scales polynomially with a number of terms in the QBM Hamiltonian (see equation 11, given that the number of terms in the QBM Hamiltonian Hw is reflected to the number of parameters nθ).
Regarding claim 14, Huijgen teaches the method of claim 1, further comprising extending, for a third stage of training, the Hamiltonian ansatz with at least one other set of operators and parameters, wherein the at least one other set of operators and parameters are optionally orthogonal (see warm-up strategy in page 2, col. 1, paragraph 4).
Regarding claim 15, Huijgen teaches the method of claim 14, further comprising initializing the parameters of the QBM with an extended Hamiltonian ansatz using optimal parameters from a previous quantum optimization loop (see warm-up strategy in page 2, col. 1, paragraph 4).
Regarding claim 16, Huijgen teaches the method of claim 1, wherein Gibbs states are used to provide samples for machine learning (see page 3, col.1, paragraph 3 , page 6, col. 2, paragraph 2).
Regarding claim 17, Huijgen teaches the method of claim 1, wherein Gibbs states used for the Quantum Boltzmann machine are prepared and sampled on the quantum computing hardware and the parameters are maintained on the classical binary computing hardware (see page 3, col.1, paragraph 3, abstract, page 2, col. 1, paragraph 2).
Regarding claim 18, the machine learning system of claim 20 is rejected under the same arts and evidence used to reject claim 1.
Regarding claim 19, Huijgen teaches the system according to claim 18, wherein the quantum computing hardware is implemented using a plurality of qubits which can each be programmatically connected to any other qubit of the plurality (see Figure 1 top, full connectivity).
Regarding claim 20, the machine learning system of claim 20 is rejected under the same arts and evidence used to reject claim 1.
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.
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Huijgen in view of Verdon (Quantum Hamiltonian-Based Models & the Variational Quantum Thermalizer Algorithm).
Regarding claim 5, Huijgen teaches the method of claim 1. Huijgen does not explicitly disclose wherein the first stage of training is performed using a mean-field (MF) model, a one-dimensional or two-dimensional geometrically local (GL) model, and/or a Gaussian Fermionic (GF) model.
In an analogous art, Verdon teaches the first stage of training is performed using a mean-field (MF) model, a one-dimensional or two-dimensional geometrically local (GL) model, and/or a Gaussian Fermionic (GF) model (discloses such a Gaussian Fermionic model (see abstract: "We use QHBMs and the VQT on Heisenberg spin systems, we apply QHBMs to learn entanglement Hamiltonians and compression codes in simulated free Bosonic systems, and finally we use the VQT to prepare thermal Fermionic Gaussian states for quantum simulation.").
It would have been obvious for a person of ordinary skill in the art, before the effective filling date of the claimed invention, to take the teachings of Verdon and apply them to Huijgen. One would be motivated as such as to maximize the utility of both classical and quantum processors.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HESHAM K ABOUZAHRA whose telephone number is (571)270-0425. The examiner can normally be reached M-F 8-5.
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, Jamie Atala can be reached at 57127227384. 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.
/HESHAM K ABOUZAHRA/ Primary Examiner, Art Unit 2486
Prosecution Insights