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 .
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 .
Response to Remarks
Claim Rejections – 35 U.S.C. 101
Applicant’s amendments have been fully considered but they are not persuasive.
Applicant argues (pg. 6) that the rejection fails to establish a prima facie case under Step 2A, Prong 1. In particular, Applicant alleges that Examiner’s Step 2A Prong 1 analysis omits steps (b) and (c) entirely. Examiner respectfully disagrees. Steps (b) and (c) of claim 1 is discussed in Step 2A Prong 2 instead, as they are instances of high-level recitations of training machine learning model to generate quantum generative model / bit string samples.
Applicant argues (pg. 7) that only steps (b) and (c) are addressed under Step 2A Prong 2. Applicant states that the Examiner’s characterization of steps (b) and (c) as mere “apply it” instructions is incorrect because they are not generic computer instructions, rather they are quantum machine learning operations. Examiner respectfully disagrees. Just because they are a quantum process does not mean that the steps amount to more than a high-level recitation of machine learning training.
The foregoing applies to all independent claims and their dependent claims.
Claim Rejections – 35 U.S.C. 103
Applicant’s amendments have been fully considered but they are only partially persuasive.
Applicant argues (pg. 9-10) that Han does not teach distinct generating and filtering steps.
Examiner agrees. Accordingly, a new reference, Martinis (“Quantum Supremacy Using a Programmable Superconducting Processor”) has been added to the rejection, as further detailed below.
Applicant argues (pg. 10-11) that Han does not teach the application of the cost function onto the filtered bitstring samples.
Examiner agrees. Accordingly, a new reference, Killoran (“Quantum computation with neutral atoms”) has been added to the rejection, as further detailed below.
Applicant argues (pg. 11-12) that Han does not teach the iterative repeating step of claim 1.
Examiner respectfully disagrees. Han, page 7, col. 1, paragraph 1, states: “As the number of training patterns increases, MPS with a fixed D_{max} will eventually fail in remembering exactly all the training patterns”. Han teaches that there are multiple iterations of training in the various models that are studied (Fig 2a, 2b in Han page 6). In each of the models, there are training data that is generated (Han [Page 6, Column 1, Paragraph 4]: “We have generated Ns = 10^6 independent samples from the learned MPS. All these samples are training images”).
The foregoing applies to all independent claims and their dependent claims.
Claim Rejections – 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: Claims 1-17 are method claims. Claims 18-20 are machine/system/product claims. Therefore, claims 1-20 are directed to either a process, machine, manufacture or composition of matter.
With respect to claim 1:
Step 2A – Prong 1:
…
(a) generating a first dataset, the first dataset comprising a plurality of bit string samples from a prior probability distribution, (mental process – a person can manually generate a dataset comprising a plurality of bit string samples from a prior probability distribution with the assistance of a pen/paper.)
…
…
(d) filtering the new bit string samples according to properties of the plurality of new bit string samples to produce a plurality of filtered bit string samples; (mental process – a person can manually filter the new bit string samples according to properties of the plurality of new bit string samples to produce a plurality of filtered bit string samples with the assistance of a pen/paper.)
(e) applying a cost function to the plurality of filtered bit string samples to produce a plurality of cost function values of the plurality of filtered bit string samples; (mental process – a person can manually apply a cost function to the plurality of filtered bit string samples to produce a plurality of cost function values of the plurality of filtered bit string samples with the assistance of a pen/paper.)
(f) evaluating the plurality of filtered bit string samples based on the plurality of cost function values of the plurality of filtered bit string samples; (mental process – a person can manually apply a cost function to the plurality of filtered bit string samples to produce a plurality of cost function values of the plurality of filtered bit string samples with the assistance of a pen/paper.)
(g) selecting a subset of the plurality of filtered bit string samples based on the evaluation; (mental process – a person can manually select a subset of the plurality of filtered bit string samples based on the evaluation with the assistance of a pen/paper.)
(h) merging the first dataset with the subset of the plurality of filtered bit string samples to generate a second dataset; (mental process – a person can manually merge the first dataset with the subset of the plurality of filtered bit string samples to generate a second dataset with the assistance of a pen/paper.)
and (g) iteratively repeating (c) through (h), wherein in each iteration the output of (h) provides the input to (c) until reaching a limiting number of iterations. (mental process – a person can manually repeating (c) through (h), wherein in each iteration the output of (h) provides the input to (c) until reaching a limiting number of iterations with the assistance of a pen/paper.)
Step 2A – Prong 2: This judicial exception is not integrated into a practical application.
A method performed by a computer system for solving combinatorial optimization problems, the computer system comprising a classical computer, the classical computer including a processor, a non-transitory computer-readable medium, and computer instructions stored in the non-transitory computer-readable medium, the computer program instructions being executable by the processor to perform the method, the method comprising: (mere instructions to apply the exception using a generic computer component – computer, processor, non-transitory computer-readable medium applies exception.)
…
(b) performing unsupervised training using the first dataset to generate a quantum generative model; (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: High level recitation of performing unsupervised training using the first dataset to generate a quantum generative model.)
(c) using the quantum generative model to generate a plurality of new bit string samples; (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: High level recitation of training the machine learning engine to generate a plurality of new bit string samples.)
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 2:
Step 2A – Prong 1:
The method of claim 1, wherein the properties of the plurality of new bit string samples include cardinality constraints. (mental process – a person can recognize that the properties of the plurality of new bit string samples include cardinality constraints.)
With respect to claim 3:
Step 2A – Prong 1:
The method of claim 1, wherein the properties of the plurality of new bit string samples include frequency of appearance. (mental process – a person can recognize that the properties of the plurality of new bit string samples include frequency of appearance.)
With respect to claim 4:
Step 2A – Prong 1:
The method of claim 1, wherein the prior probability distribution comprises initial observations and cost function values. (mental process – a person can recognize that the prior probability distribution comprises initial observations and cost function values.)
With respect to claim 5:
Step 2A – Prong 1:
The method of claim 4, further comprising drawing the initial observations from randomly selected data elements in the first dataset. (mental process – a person can recognize that drawing the initial observations from randomly selected data elements in the first dataset.)
With respect to claim 6:
Step 2A – Prong 1:
The method of claim 1, wherein (b) comprises using matrix product states (MPS) to generate the quantum generative model. (mental process – a person can recognize that (b) comprises using matrix product states (MPS) to generate the quantum generative model.)
With respect to claim 7:
Step 2A – Prong 1:
The method of claim 1, wherein the quantum generative model is implemented as a tensor network (TN). (mental process – a person can recognize that the quantum generative model is implemented as a tensor network (TN).)
With respect to claim 8:
Step 2A – Prong 1:
The method of claim 1, wherein the quantum generative model comprises a generative adversarial network (GAN). (mental process – a person can recognize that the quantum generative model comprises a generative adversarial network (GAN).)
With respect to claim 9:
Step 2A – Prong 1:
The method of claim 1, wherein the evaluating comprises evaluating the plurality of new bit string samples based on minimizing cost function values. (mental process – a person can recognize that the evaluating comprises evaluating the plurality of new bit string samples based on minimizing cost function values.)
With respect to claim 10:
Step 2A – Prong 1:
The method of claim 1, wherein the method is practiced in a stand-alone mode. (mental process – a person can recognize that the method is practiced in a stand-alone mode.)
With respect to claim 11:
Step 2A – Prong 1:
The method of claim 10, wherein the required number of cost function evaluations is smaller than that of classical optimizers. (mental process – a person can recognize that the required number of cost function evaluations is smaller than that of classical optimizers.)
With respect to claim 12:
Step 2A – Prong 1:
The method of claim 1, wherein (a) comprises receiving the first dataset from an output of a first optimizer, and wherein the method boosts performance of the first optimizer. (mental process – a person can recognize that (a) comprises receiving the first dataset from an output of a first optimizer, and wherein the method boosts performance of the first optimizer.)
With respect to claim 13:
Step 2A – Prong 1:
The method of claim 12, wherein the method achieves lower minima of the cost function than the first optimizer. (mental process – a person can recognize that the method achieves lower minima of the cost function than the first optimizer.)
With respect to claim 14:
Step 2A – Prong 1:
The method of claim 13, wherein the first optimizer comprises a classical optimizer. (mental process – a person can recognize that the first optimizer comprises a classical optimizer.)
With respect to claim 15:
Step 2A – Prong 1:
The method of claim 1, wherein the computer system further comprises a quantum computer, the quantum computer comprising a plurality of qubits. (mental process – a person can recognize that the computer system further comprises a quantum computer, the quantum computer comprising a plurality of qubits.)
With respect to claim 16:
Step 2A – Prong 1:
The method of claim 15, wherein performing unsupervised training using the first dataset to generate the quantum generative model comprises performing the unsupervised training on the quantum computer. (mental process – a person can recognize that performing unsupervised training using the first dataset to generate the quantum generative model comprises performing the unsupervised training on the quantum computer.)
With respect to claim 17:
Step 2A – Prong 1:
The method of claim 15, wherein the quantum generative model comprises a quantum-assisted generative adversarial network (qa-GAN). (mental process – a person can recognize that the quantum generative model comprises a quantum-assisted generative adversarial network (qa-GAN).)
Claims 18, 19, 20 are rejected on the same grounds under 35 U.S.C. 101 as claims 1, 4, 2 as they are substantially similar, respectively. Mutatis mutandis.
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 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Han et al. (“Unsupervised Generative Modeling Using Matrix Product States”) hereinafter known as Han in view of Mazzola et al. (US 20220188679 A1) hereinafter known as Mazzola in view of Martinis (“Quantum Supremacy Using a Programmable Superconducting Processor”) hereinafter known as Martinis in view of Killoran (“Quantum computation with neutral atoms”) hereinafter known as Killoran.
Regarding independent claim 1, Han teaches:
…
(a) generating a first dataset, the first dataset comprising a plurality of bit string samples from a prior probability distribution, (Han [Page 2, Column 2, Paragraph 1]: “The goal of unsupervised generative modeling is to model the joint probability distribution of given data. With the trained model, one can then generate new samples from the learned probability distribution. Generative modeling finds wide applications such as dimensional reduction, feature detection, clustering, and recommendation systems. In this paper, we consider a data set T consisting of binary strings v ∈ V = {0, 1} ^ ⊗N, which are potentially repeated and can be mapped to basis vectors of a Hilbert space of dimension 2^N” Han teaches that the first dataset is generated and is comprised of binary strings from a probability distribution.)
(b) performing unsupervised training using the first dataset to generate a quantum generative model; (Han [Page 2, Column 2, Paragraph 1]: “The goal of unsupervised generative modeling is to model the joint probability distribution of given data. With the trained model, one can then generate new samples from the learned probability distribution.” Han teaches that unsupervised training is used on the dataset to generate a quantum generative model.)
…
(d) filtering the new bit string samples according to properties of the plurality of new bit string samples to produce a plurality of filtered bit string samples; (Han [Page 4, Column 1, Paragraph 6]: “Combining Eqs. (7) and (9), one can compute the conditional probability (6) and sample the bit vk−1 accordingly” Han teaches that the bitstrings are constructed using a probability that gives whether a bit at a given index is 1. This is essentially filtering an all 0 bitstring by setting the bits in certain indexes to 1.)
…
…
…
(h) merging the first dataset with the subset of the plurality of filtered bit string samples to generate a second dataset; (Han [Page 6, Column 1, Paragraph 4]: “We have generated Ns = 10^6 independent samples from the learned MPS. All these samples are training images” Han teaches that the samples that are generated are all merged and used as training data.)
and (g) iteratively repeating (c) through (h), wherein in each iteration the output of (h) provides the input to (c) until reaching a limiting number of iterations. (Han [Page 5, Column 2, Paragraph 6]: “After being trained over four loops of batch gradient descent training, the cost function converges to its minimum value” Han teaches that the process is iterated until the ending number of iterations is reached, which in the example is four.)
Han does not explicitly teach:
A method performed by a computer system for solving combinatorial optimization problems, the computer system comprising a classical computer, the classical computer including a processor, a non-transitory computer-readable medium, and computer instructions stored in the non-transitory computer-readable medium, the computer program instructions being executable by the processor to perform the method, the method comprising:
However, Mazzola teaches:
A method performed by a computer system for solving combinatorial optimization problems, the computer system comprising a classical computer, the classical computer including a processor, a non-transitory computer-readable medium, and computer instructions stored in the non-transitory computer-readable medium, the computer program instructions being executable by the processor to perform the method, the method comprising: (Mazzola [¶ 0005]: “a computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor” Mazzola teaches that the instructions are stored in a non-transitory medium in a computing environment accessible to a processor.)
Han and Mazzola are in the same field of endeavor as the present invention, as the references are directed to using quantum computing methods to sample data using a quantum generative model. It would have been obvious, before the effective filing date of the claimed invention, to a person of ordinary skill in the art, to combine generating and accessing data using a cost function as taught in Han with using a classical optimizer set up the data for the subsequent quantum classifier as taught in Mazzola. Mazzola provides this additional functionality. As such, it would have been obvious to one of ordinary skill in the art to modify the teachings of Han to include teachings of Mazzola because the combination would allow for the technique of quantum computing to be done on a classical computer via emulation or a quantum computer. This has the potential benefit of using a classical computer to gain the increased of quantum computing techniques.
Han and Mazzola do not explicitly teach:
(c) using the quantum generative model to generate a plurality of new bit string samples;
However, Martinis teaches:
(c) using the quantum generative model to generate a plurality of new bit string samples; (Martinis [Page 3, Paragraph 1]: “Each run of a random quantum circuit on a quantum computer produces a bitstring, for example 0000101.” Martinis teaches that the run of the quantum circuit generates a bitstring. This is a model that outputs a bitstring that can be used for training data.)
Martinis is in the same field as the present invention, since it is directed to generating bitstrings using quantum circuits/models. It would have been obvious, before the effective filing date of the claimed invention, to a person of ordinary skill in the art, to combine filtering the bitstring as taught in Han as modified by Mazzola with generating a new bitstring as taught in Martinis. Martinis provides this additional functionality. As such, it would have been obvious to one of ordinary skill in the art to modify the teachings of Han as modified by Mazzola to include teachings of Martinis because the combination would allow for the generation of bitstrings to be filtered afterward. This has the potential benefit of cleaning up the data and making it more amenable for training use.
Han, Mazzola, and Martinis do not explicitly teach:
(e) applying a cost function to the plurality of filtered bit string samples to produce a plurality of cost function values of the plurality of filtered bit string samples;
(f) evaluating the plurality of filtered bit string samples based on the plurality of cost function values of the plurality of filtered bit string samples;
(g) selecting a subset of the plurality of filtered bit string samples based on the evaluation;
However, Killoran teaches:
(e) applying a cost function to the plurality of filtered bit string samples to produce a plurality of cost function values of the plurality of filtered bit string samples; (Killoran [Page 8, Paragraph 1]: “With this goal in mind, we can create a basic cost function. This cost function randomly samples possible 4-bit input bitstrings, and compares the circuit’s output with the value of the first bit. The other bits can be thought of as noise that we don’t want our model to learn.” Killoran teaches that the bitstrings are individually applied to the cost function.)
(f) evaluating the plurality of filtered bit string samples based on the plurality of cost function values of the plurality of filtered bit string samples; (Killoran [Page 8, Paragraph 1]: “This cost function randomly samples possible 4-bit input bitstrings, and compares the circuit’s output with the value of the first bit. The other bits can be thought of as noise that we don’t want our model to learn.” Killoran teaches that after the bitstrings are individually applied to the cost function, they are compared with the value of the first bit, meaning that they are evaluated based on the cost function value.)
(g) selecting a subset of the plurality of filtered bit string samples based on the evaluation; (Killoran [Page 8, Paragraph 2]: “The circuit has learned to transfer the state of the first qubit to the state of the last qubit, while ignoring the state of all other input qubits” Based on the evaluation of the bitstring on the cost function, the state of the index of the bits are evaluated and selected (while the rest are ignored).)
Killoran is in the same field as the present invention, since it is directed to applying a cost function during training to an individual bitstring. It would have been obvious, before the effective filing date of the claimed invention, to a person of ordinary skill in the art, to combine filtering the bitstring as taught in Han as modified by Mazzola as modified by Martinis with applying a cost function to the bitstring as taught in Killoran. Killoran provides this additional functionality. As such, it would have been obvious to one of ordinary skill in the art to modify the teachings of Han as modified by Mazzola as modified by Martinis to include teachings of Killoran because the combination would allow for the fit of the bitstring to be assessed by applying it to the cost function. This has the potential benefit of determining which bitstrings are best for training use, allowing keeping the best bitstrings while discarding the others.
Regarding dependent claim 2, Han and Mazzola teach:
The method of claim 1, wherein the properties of the plurality of new bit string samples include cardinality constraints. (Han [Page 7, Column 1, Paragraph 2]: “a larger Dmax enables MPS to remember exactly more patterns and produce smaller L with the number of patterns abs(T) fixed” Han teaches that the constraints include one where the size of the patterns is the number of samples in the training data.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 3, Han and Mazzola teach:
The method of claim 1, wherein the properties of the plurality of new bit string samples include frequency of appearance. (Han [Page 4, Column 1, Paragraph 6]: “Combining Eqs. (7) and (9), one can compute the conditional probability (6) and sample the bit vk−1 accordingly” Han teaches that the bitstrings are constructed using a probability that gives whether a bit at a given index is 1. This probability function is equivalent to the frequency of appearance for a given index in a bitstring.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 4, Han and Mazzola teach:
The method of claim 1, wherein the prior probability distribution comprises initial observations and cost function values. (Han [Page 2, Column 2, Paragraph 1]: “The goal of unsupervised generative modeling is to model the joint probability distribution of given data. With the trained model, one can then generate new samples from the learned probability distribution. Generative modeling finds wide applications such as dimensional reduction, feature detection, clustering, and recommendation systems. In this paper, we consider a data set T consisting of binary strings v ∈ V = {0, 1} ^ ⊗N, which are potentially repeated and can be mapped to basis vectors of a Hilbert space of dimension 2^N” Han teaches that the first dataset is generated and is comprised of binary strings from a probability distribution. These have initial observations, as they can be mapped to a Hilbert space of dimension 2^N, and thus can be input to the cost function to get a negative log likelihood value.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 5, Han and Mazzola teach:
The method of claim 4, further comprising drawing the initial observations from randomly selected data elements in the first dataset. (Han [Page 4, Column 1, Paragraph 6]: “Combining Eqs. (7) and (9), one can compute the conditional probability (6) and sample the bit vk−1 accordingly” Han teaches that the bitstrings are constructed using a probability that gives whether a bit at a given index is 1. The probability of observing a data element is equal to this distribution.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 6, Han and Mazzola teach:
The method of claim 1, wherein (b) comprises using matrix product states (MPS) to generate the quantum generative model. (Han [Page 2, Column 1, Paragraph 2]: “In particular, the matrix product state (MPS) is a kind of TN where the tensors are arranged in a one-dimensional geometry” Han teaches that the matrix product states are used to generate the quantum generative model.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 7, Han and Mazzola teach:
The method of claim 1, wherein the quantum generative model is implemented as a tensor network (TN). (Han [Page 2, Column 1, Paragraph 2]: “In particular, the matrix product state (MPS) is a kind of TN where the tensors are arranged in a one-dimensional geometry” Han teaches that the quantum generative model is implemented as a tensor network.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 8, Han and Mazzola teach:
The method of claim 1, wherein the quantum generative model comprises a generative adversarial network (GAN). (Mazzola [¶ 0083]: “An alternative method to load the path distribution can involve use of a quantum Generative Adversarial Network (qGAN).” Mazzola teaches that the quantum generative model comprises a quantum Generative Adversarial Network.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 9, Han and Mazzola teach:
The method of claim 1, wherein the evaluating comprises evaluating the plurality of new bit string samples based on minimizing cost function values. (Han [Page 3, Column 1, Paragraph 6]: “Minimizing the NLL reduces the dissimilarity between the model probability distribution P(v) and the empirical distribution defined by the training set.” Han teaches that the negative log likelihood is the cost function and is applied to the filtered bitstring samples.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 10, Han and Mazzola teach:
The method of claim 1, wherein the method is practiced in a stand-alone mode. (Han [Page 5, Column 2, Paragraph 6]: “After being trained over four loops of batch gradient descent training, the cost function converges to its minimum value” Han teaches that the process is iterated until the ending number of iterations is reached, which in the example is four. This process of self-iterating shows how it can operate in a stand-alone way.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 11, Han and Mazzola teach:
The method of claim 10, wherein the required number of cost function evaluations is smaller than that of classical optimizers. (Han [Page 5, Column 2, Paragraph 6]: “After being trained over four loops of batch gradient descent training, the cost function converges to its minimum value” Han teaches that the process is iterated until the ending number of iterations is reached, which in the example is four. Since it is an improvement in efficiency to existing classical systems, in many cases, there will be fewer evaluations needed.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 12, Han and Mazzola teach:
The method of claim 1, wherein (a) comprises receiving the first dataset from an output of a first optimizer, and wherein the method boosts performance of the first optimizer. (Mazzola [¶ 0362]: “Since quantum resource estimation system 102 can use classical emulation of the quantum circuits the only source of error in the optimizations is originated from the classical optimizer.” Mazzola teaches that when the emulation of the quantum resources is done on the classical computer, then necessarily there is a first optimizer, that is a classical optimizer.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 13, Han and Mazzola teach:
The method of claim 12, wherein the method achieves lower minima of the cost function than the first optimizer. (Mazzola [¶ 0362]: “Since quantum resource estimation system 102 can use classical emulation of the quantum circuits the only source of error in the optimizations is originated from the classical optimizer.” Mazzola teaches that when the emulation of the quantum resources is done on the classical computer, then necessarily there is a first optimizer, that is a classical optimizer. This optimizer is also less optimized than the quantum optimizer, and thus has a higher minimum of the cost function.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 14, Han and Mazzola teach:
The method of claim 13, wherein the first optimizer comprises a classical optimizer. (Mazzola [¶ 0362]: “Since quantum resource estimation system 102 can use classical emulation of the quantum circuits the only source of error in the optimizations is originated from the classical optimizer.” Mazzola teaches that when the emulation of the quantum resources is done on the classical computer, then necessarily there is a first optimizer, that is a classical optimizer.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 15, Han and Mazzola teach:
The method of claim 1, wherein the computer system further comprises a quantum computer, the quantum computer comprising a plurality of qubits. (Mazzola [¶ 0356]: “The initial state, defined on a n qubits register, which quantum resource estimation system” Mazzola teaches that the quantum computer comprises of qubits.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 16, Han and Mazzola teach:
The method of claim 15, wherein performing unsupervised training using the first dataset to generate the quantum generative model comprises performing the unsupervised training on the quantum computer. (Han [Page 2, Column 1, Paragraph 3]: “building on the connection between unsupervised generative modeling and quantum physics, we employ MPS as a model to learn the probability distribution of given data with an algorithm” Han teaches that the unsupervised learning is done on the quantum computer, as evidenced by the use of the MPS.)
The reasons to combine are substantially similar to those of claim 1.
Regarding dependent claim 17, Han and Mazzola teach:
The method of claim 15, wherein the quantum generative model comprises a quantum-assisted generative adversarial network (qa-GAN). (Mazzola [¶ 0083]: “An alternative method to load the path distribution can involve use of a quantum Generative Adversarial Network (qGAN).” Mazzola teaches that the quantum generative model comprises a quantum Generative Adversarial Network.)
The reasons to combine are substantially similar to those of claim 1.
Claims 18, 19, 20 are rejected on the same grounds under 35 U.S.C. 103 as claims 1, 4, 2 as they are substantially similar, respectively. Mutatis mutandis.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KYU HYUNG HAN whose telephone number is (703) 756-5529. The examiner can normally be reached on MF 9-5.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Alexey Shmatov can be reached on (571) 270-3428. 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.
/Kyu Hyung Han/
Examiner
Art Unit 2123
/ALEXEY SHMATOV/Supervisory Patent Examiner, Art Unit 2123
Prosecution Insights