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 .
Response to Amendment
The amendment filed 08/10/2026 has been entered. Claims 1-20 remain pending in the application.
Response to Arguments
Applicant’s arguments, filed 08/10/2026, with respect to the rejections of the claims under 103 have been fully considered and are persuasive because of the amendments. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of Penghua et al. (A Quantum Self-Organizing Mapping Neural Network) in view of Corcoles-Gonzalez et al. (US Pub. 2019/0378025) in view of Kohonen (The Self-Organizing Map) in view of Asan et al. (An Introduction to Self-Organizing Maps, Chapter 14) and further in view of Kshirsagar et al. (US Pub. 2025/0117677).
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-10 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor, or for pre-AIA the applicant regards as the invention.
Claim 1 is rejected for reciting a limitation of “Output of inner part …”. There is insufficient antecedent basis for this limitation in the claim. For examination purposes, limitation “Output of inner part …” is interpreted as “Output of inner product …”.
Claims 2-10 are rejected as being dependent upon a rejected base claim, namely claim 1.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 1-2, 5-6, 9, 12, 14, 17 and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Penghua et al. (A Quantum Self-Organizing Mapping Neural Network) in view of Corcoles-Gonzalez et al. (US Pub. 2019/0378025) in view of Kohonen (The Self-Organizing Map - Applicant provided NPL in the IDS) in view of Asan et al. (An Introduction to Self-Organizing Maps, Chapter 14) and further in view of Kshirsagar et al. (US Pub. 2025/0117677).
As per claim 1, Penghua teaches a system, comprising:
an estimating component that executes a set of quantum circuits resulting in measured outputs that define estimated inner products overlap values between a sample quantum state, of a set of quantum data, and a set of weight vectors, associated with quantum-state representations of a set of weight vectors associated with neurons of a lattice space of a neural network [Fig. 1, abstract, “A new quantum SOM (self-organizing map) neural model with elastic neighborhood radius is proposed. The new model objects the real input data into the quantum initial states. Through the operation of the quantum gate in the qubit neuron, the quantum initial states are converted into the quantum intermediate states, then into the quantum excited sates. These excited sates, being connected with the quantized weights, are perceived by the competitive-layer neurons. According to a new quantized competitive learning algorithm, the input data can be orderly topology mapped”; pages 3264-3265, section 2.1, “The topology of this new neural model is shown in Fig. 1. In this neural model, the quantized input data are regarded as the quantum initial states. These initial states are inspired to be the quantum excited states using the mapping of the qubit neuron … we consider a n-dimensional real input vector x (k) = [xi(k), X2 (k) .. · Xn (k)]T … It can be converted into the quantum state … The quantum initial states, described as equation (2), are inspired into quantum transient sates using the rotating operation of the qubit neuron, then are inspired into the quantum excited states … These quantum excited states are connected with the quantized weights of the neurons in the competitive layer … the input-output relationship of the corresponding competitive-layer neurons can be described as
PNG
media_image1.png
52
286
media_image1.png
Greyscale
It is noteworthy that, yj(k) is not the actual output of the network, but the index value of the winning neuron in the competitive layer or the weight vector connected with the quantum excited state; It can be seen that the output y defining the inner product of the quantum state and the weights of the neurons of a lattice space of a neural network as shown in Fig. 1 and equation 9”], the estimated overlap values being derived from the measured outputs of the executed set of quantum circuits [page 3265, equation 9 shows the inner product overlap values is obtained from the output y
PNG
media_image1.png
52
286
media_image1.png
Greyscale
]; and
a determining component that, based on the estimated inner products measurement- derived estimated overlap values, identifies a distance, of the neurons of the lattice space, to the sample quantum state within a Hilbert space of the neural network, the Hilbert space comprising the set of quantum data and the lattice space [Fig. 1, abstract, “A new quantum SOM (self-organizing map) neural model with elastic neighborhood radius is proposed. The new model objects the real input data into the quantum initial states. Through the operation of the quantum gate in the qubit neuron, the quantum initial states are converted into the quantum intermediate states, then into the quantum excited sates. These excited sates, being connected with the quantized weights, are perceived by the competitive-layer neurons. According to a new quantized competitive learning algorithm, the input data can be orderly topology mapped. The elastic neighborhood radius, in the new learning algorithm, is defined by both of the similarity and the distance between quantized weights and quantum excited states”; page 3264, Introduction, “The self-organizing map (SOM) [1,2] is an unsupervised neural network approach that reduces a high-dimensional data set to a representative and compact two-dimensional grid; pages 3264-3265, section 2.1, “The topology of this new neural model is shown in Fig. 1. In this neural model, the quantized input data are regarded as the quantum initial states. These initial states are inspired to be the quantum excited states using the mapping of the qubit neuron … we consider a n-dimensional real input vector x (k)”; Examiner interprets the high-dimensional or n-dimensional space as the Hilbert space which comprising quantum data and lattice space as shown in Fig. 1], wherein the quantum-state representations of the set of weight vectors reside within the Hilbert space [Fig. 1 shows the weight vectors associated with the quantum states within the Hilbert space].
Penghua does not explicitly teach
a memory that stores computer executable components; and
a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:
executes a set of quantum circuits at a quantum processor;
based on the estimated inner products measurement- derived estimated overlap values, identifies a best matching neuron of the lattice space having a closest distance, of the neurons of the lattice space, to the sample quantum state.
Output of inner part determines match.
Corcoles-Gonzalez teaches
a memory that stores computer executable components [Fig. 1, memory 110]; and
a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise [Fig. 1, paragraph 0005, “a processor that executes computer executable components stored in a memory; a quantum feature map circuit component that estimates a kernel associated with a feature map”]:
executes a set of quantum circuits at a quantum processor [Fig. 1, paragraph 0005, “a processor that executes computer executable components stored in a memory; a quantum feature map circuit component that estimates a kernel associated with a feature map”];
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include executing a set of quantum circuits at a quantum processor of Corcoles-Gonzalez. Doing so would help estimating a kernel associated with a feature map and performing a classification using the estimated kernel (Corcoles-Gonzalez, 0005).
Penghua and Corcoles-Gonzalez do not explicitly teach
based on the estimated inner products measurement- derived estimated overlap values, identifies a best matching neuron of the lattice space having a closest distance, of the neurons of the lattice space, to the sample quantum state.
Output of inner part determines match.
Kohonen teaches
based on the estimated inner products measurement- derived estimated overlap values, identifies a best matching neuron [page 1467, Col. 1, lines 3-7, “The simplest analytical measure for the match of x with the mi may be the inner product xmi”].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include identifying a best matching neuron based on the estimated inner products of Cohonen. Doing so would help obtaining the competitive learning rule of quantum SOM neural network (Penghua, 3265, section 2.2).
Penghua, Corcoles-Gonzalez and Kohonen do not explicitly teach
identifies a best matching neuron of the lattice space having a closest distance, of the neurons of the lattice space, to the sample quantum state.
Output of inner part determines match.
Asan teaches
identifies a best matching neuron of the lattice space having a closest distance, of the neurons of the lattice space, to the sample quantum state [page 300, last paragraph, “The output nodes (neurons) in a self-organizing map compete with each other to best represent the particular input sample. The success of representation is measured using a discriminant function, where an input vector is compared with the weight vector of each output node. The particular node with its connection weights most similar to the input sample is declared winner of the competition. There are a number of different functions to determine the winner, i.e. the best-matching unit (BMU) on the map. The most commonly used one is the Euclidean distance”; page 306, section 14.2.2.2 Similarity matching, “learning in SOM does not rely on predefined target outcomes that would guide the process. The output nodes compete among themselves to become activated. Only the node whose weight vector is most similar to the input vector will be activated and declared as the winner [20]. To find this best matching node, the distances between an input data (x) and all the weight vectors (w;) of the SOM are computed using different measurement methods, such as … Manhattan distance … Euclidean distance”].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include identifying a best matching neuron of the lattice space having a closest distance of the neurons of the lattice space to the input vector of Asan. Doing so would help adjusting the weight vector of the best matching neuron and its neighborhood until the change of the weights is less than a threshold (Asan, pages 311-312, section 14.2.2.4).
Penghua, Corcoles-Gonzalez, Kohonen and Asan do not explicitly teach
Output of inner part determines match.
Kshirsagar teaches
Output of inner part determines match [paragraph 0054, “when the first and second quantum states are represented as first and second vectors, respectively, the first quantum state approximates the second quantum state when an inner product between the first and second vectors (called the "fidelity" between the two quantum states) is greater than a predefined amount (typically labeled E). In this example, the fidelity quantifies how "close" or "similar" the first and second quantum states are to each other].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include output of inner part determines match of Kshirsagar. Doing so would help obtaining the competitive learning rule of quantum SOM neural network (Penghua, 3265, section 2.2).
As per claim 2, Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar teach the system of claim 1.
Penghua further teaches
an updating component that, based on the estimation of the inner products measurement- derived estimated overlap values, updates a subset of the weight vectors [pages 3265-3266, section 2.2, “Through the weight adjustment described by equation (18)”];
Asan further teaches
an updating component that, based on employing a quantum kernel, updates a subset of the weight vectors of corresponding subset of the neurons of the lattice space that have a selected proximity to the best matching neuron [page 301, (iii), “The weight vectors of the winner and its neighboring units in the map are adjusted”; page 309, 1st paragraph, “Finally, after defining the neighborhood set or kernel around the winning node c [16], a neighborhood function is chosen to update the weight vectors of the respective nodes”; page 310, 2nd paragraph, “the winning node was the first output node (c = 1) that had the minimum Euclidean distance. Now the weight vectors of this best matching node and its neighborhood will be adjusted so that they become more similar to the current input sample x(0)”],
wherein the subset of adjustable weights is disposed within a selected threshold distance of the best matching neuron [page 309, 1st paragraph, “Finally, after defining the neighborhood set or kernel around the winning node c [16], a neighborhood function is chosen to update the weight vectors of the respective nodes”; page 310, 2nd paragraph, “the winning node was the first output node (c = 1) that had the minimum Euclidean distance. Now the weight vectors of this best matching node and its neighborhood will be adjusted so that they become more similar to the current input sample x(0)”; examiner interprets the process of adjusting the weights of the winning node c which has the minimum Euclidean distance and its neighbor nodes as disposing the adjusted weights with a threshold distance of node c].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include updating a subset of the weight vectors of corresponding subset of the neurons of the lattice space that have a selected proximity to the best matching neuron, wherein the subset of adjustable weights is disposed within a selected threshold distance of the best matching neuron of Asan. Doing so would help the relevant nodes become more similar to the input sample (Asan, page 301, (iii)).
As per claim 5, Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar teach the system of claim 1.
Kshirsagar further teaches
a fidelity component that, based on the estimation of the inner products measurement- derived estimated overlap values resulting in transitional probabilities between the sample of quantum data and the set of weight vectors the quantum-state representations of the set of weight vectors, generates fidelities between the quantum state sample quantum state and other quantum states of the set of quantum data, wherein the determining component employs the fidelities in a fidelity-based metric to identify the best matching neuron [paragraph 0054, “when the first and second quantum states are represented as first and second vectors, respectively, the first quantum state approximates the second quantum state when an inner product between the first and second vectors (called the "fidelity" between the two quantum states) is greater than a predefined amount (typically labeled E). In this example, the fidelity quantifies how "close" or "similar" the first and second quantum states are to each other. The fidelity represents a probability that a measurement of the first quantum state will give the same result as if the measurement were performed on the second quantum state”].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include generating fidelities between the quantum state and other quantum states of the set of quantum data, wherein the determining component employs the fidelities in a fidelity-based metric to identify the best matching neuron of Kshirsagar. Doing so would help adjusting the weight vector of the best matching neuron and its neighborhood until the change of the weights is less than a threshold (Penghua (as modified), Asan, pages 311-312, section 14.2.2.4).
As per claim 6, Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar teach the system of claim 1.
Penghua teaches
the Hilbert space comprises a first input layer comprising a first set of neurons having the sample quantum state and other quantum states of the set of quantum data mapped to the first set of neurons [page 3264, Introduction, “The self-organizing map (SOM) [1,2] is an unsupervised neural network approach that reduces a high-dimensional data set to a representative and compact two-dimensional grid; pages 3264-3265, section 2.1, “The topology of this new neural model is shown in Fig. 1. In this neural model, the quantized input data are regarded as the quantum initial states. These initial states are inspired to be the quantum excited states using the mapping of the qubit neuron … we consider a n-dimensional real input vector x (k)”; Examiner interprets the high-dimensional or n-dimensional space as the Hilbert space which comprising quantum data and lattice space as shown in Fig. 1; Fig. 1 shows a n-dimensional space (Hilbert space) comprising a first input layer comprising a first set of neurons having the sample quantum state and other quantum states mapped to the first set of neurons], and
Corcoles-Gonzalez teaches
wherein the Hilbert space further comprises a second output layer comprising the neurons of the lattice space [page 303, Fig. 14.1 shows the high-dimensional space comprising the output layer comprising the neurons of the lattice space].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include the Hilbert space further comprises a second output layer comprising the neurons of the lattice space of Corcoles-Gonzalez. Doing so would help reducing a high-dimensional input space into a simpler low-dimensional output space (Corcoles-Gonzalez, page 300, 2nd paragraph).
As per claim 9, Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar teach the system of claim 1.
Penghua further teaches
an encoding component that encodes the set of quantum data as quantum states at the neural network employing quantum feature mapping [pages 3264-3265, section 2.1, “The topology of this new neural model is shown in Fig. 1. In this neural model, the quantized input data are regarded as the quantum initial states. These initial states are inspired to be the quantum excited states using the mapping of the qubit neuron … we consider a n-dimensional real input vector x (k) = [xi(k), X2 (k) .. · Xn (k)]T … It can be converted into the quantum state].
Claim 12 is rejected by the same reason as of claim 2, since these claims recite the similar limitations.
Claim 14 is rejected by the same reason as of claim 5, since these claims recite the similar limitations.
Claim 17 is rejected by the same reason as of claim 2, since these claims recite the similar limitations.
Claim 19 is rejected by the same reason as of claim 5, since these claims recite the similar limitations.
Claim 20 is rejected by the same reason as of claim 6, since these claims recite the similar limitations.
Claims 3, 13 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Penghua et al. in view of Corcoles-Gonzalez et al. in view of Kohonen in view of Asan et al. in view of Kshirsagar et al. and further in view of Wierichs et al. (General parameter-shift rules for quantum gradients – IDS provided NPL).
As per claim 3, Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar teach the system of claim 2.
Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar do not teach
determining, by the updating component, gradients of kernel functions employing parameter shift rules and eigenvalues of Pauli gates corresponding to the gradients.
Wierichs teaches
determining, by the updating component, gradients of kernel functions employing parameter shift rules and eigenvalues of Pauli gates corresponding to the gradients [page 1, Col. 2, 2nd paragraph, “The so-called parameter-shift rule [16, 23, 38, 39] can be used to estimate the gradient for these optimization techniques … this two term parameter-shift rule is restricted to gates with two distinct eigenvalues, potentially requiring expensive decompositions in order to compute hardware compatible quantum gradients”; page 6, Col. 2, 1st paragraph, “As rotation gates about any multiqubit Pauli word satisfy the original parameter-shift rule as well”; page 7, section 3.6, “we will apply the stochastic parameter-shift rule to our general shift rule … the stochastic parameter-shift rule for gates of the form UF(x) = exp(i(xG + F)) (32)
where G is a Hermitian operator with eigenvalues ±1 (so that G2 = 1), e.g., a Pauli word”; page 15, section D.2, “assume the parametrized quantum circuit to consist of Pauli rotation gates exclusively. In the spirit of the generalized function reconstruction and parameter-shift rule, we would like to relax this assumption and generalize the QAD model … The first way to generalize QAD is to compute the gradient and Hessian with the generalized parameter shift rule Eq. (24) and the shift rule for Hessian entries Eq. (36)].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include determining gradients of kernel functions employing parameter shift rules and eigenvalues of Pauli gates corresponding to the gradients of Wierichs. Doing so would help optimizing the performance of a predictive model.
Claim 13 is rejected by the same reason as of claim 3, since these claims recite the similar limitations.
Claim 18 is rejected by the same reason as of claim 3, since these claims recite the similar limitations.
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Penghua et al. in view of Corcoles-Gonzalez et al. in view of Kohonen in view of Asan et al. in view of Kshirsagar et al. and further in view of Winckel et al. (US pub. 2006/0009956).
As per claim 4, Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar teach the system of claim 1.
Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar do not teach
the inner products represent transition probabilities between the sample quantum state and states corresponding to the set of weight vectors.
Winckel teaches
the inner products represent transition probabilities between the sample quantum state and states corresponding to the set of weight vectors [paragraph 0070, “in quantum mechanics, one is interested in computing inner products such as a transition probability”; Since Penghua in Fig. 1, equation 9, teaches the inner products of the quantum states and the weight vectors of the neurons, and Winckel teaches the inner products represent transition probabilities, therefore, the combination of Penghua and Winckel teaches the above claim limitation].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include the inner products represent transition probabilities of Winckel. Doing so would help determining if a solution is resolved (Winckel, 0070).
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Penghua et al. in view of Corcoles-Gonzalez et al. in view of Kohonen in view of Asan et al. in view of Kshirsagar et al. and further in view of Lazarev et al. (Quantum assisted unsupervised data clustering on the basis of neural networks).
As per claim 8, Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar teach the system of claim 1.
Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar do not explicitly teach
an inference component that, for an inference sample quantum state, based on the neural network having been trained on the set of quantum data, directs the estimating component and determining component to perform the respective executing and identifying, resulting in identification of an inferred best matching neuron to the inference sample quantum state.
Lazarev teaches
an inference component that, for an inference sample quantum state, based on the neural network having been trained on the set of quantum data, directs the estimating component and determining component to perform the respective executing and identifying, resulting in identification of an inferred best matching neuron to the inference sample quantum state [Figs. 1-2, pages 2-3, section 2, “we can train our SOFM adjusting the components through the following learning process: 1) select an input vector randomly from the set of all input vectors; 2) find a cluster vector which is closest to the input vector; 3) adjust the weights of the winning node in such a way that it becomes even closer to the input vector; 4) repeat this process many iterations until it converges. Computationally the algorithm consists of two basic procedures, selecting a winning cluster vector and updating its weights (Fig. 1). Usually the winning cluster vector is selected based on the Euclidean distance between an input vector and the cluster vectors”].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include directing the components to perform the respective executing and identifying, resulting in identification of an inferred best matching neuron to the inference sample quantum state of Lazarev. Doing so would help training the self-organizing feature map to update the components (Lazarev, page 2, section 2).
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Penghua et al. in view of Corcoles-Gonzalez et al. in view of Kohonen in view of Asan et al. in view of Kshirsagar et al. and further in view of Jiang et al. (US Pub. 2024/0370754).
As per claim 10, Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar teach the system of claim 1.
Penghua, Corcoles-Gonzalez, Kohonen, Asan and Kshirsagar do not teach
the quantum data is obtained from quantum simulation or from one or more quantum transducers.
Jiang teaches
the quantum data is obtained from quantum simulation or from one or more quantum transducers [paragraph 0077, “provide transducers to transform different types of quantum data”].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include the quantum data is obtained from quantum simulation or from one or more quantum transducers of Jiang. Doing so would help performing the quantum simulation algorithms (Jiang, 0077).
Claims 11 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Penghua et al. (A Quantum Self-Organizing Mapping Neural Network) in view of Corcoles-Gonzalez et al. (US Pub. 2019/0378025) in view of Kohonen (The Self-Organizing Map) and further in view of Asan et al. (An Introduction to Self-Organizing Maps, Chapter 14).
As per claim 11, Penghua teaches a computer-implemented method, comprising:
executing, a set of quantum circuits resulting in measured outputs that define estimated inner products overlap values between a sample quantum state, of a set of quantum data, and a set of weight vectors, associated with quantum-state representations of a set of weight vectors associated with neurons of a lattice space of a neural network [Fig. 1, abstract, “A new quantum SOM (self-organizing map) neural model with elastic neighborhood radius is proposed. The new model objects the real input data into the quantum initial states. Through the operation of the quantum gate in the qubit neuron, the quantum initial states are converted into the quantum intermediate states, then into the quantum excited sates. These excited sates, being connected with the quantized weights, are perceived by the competitive-layer neurons. According to a new quantized competitive learning algorithm, the input data can be orderly topology mapped”; pages 3264-3265, section 2.1, “The topology of this new neural model is shown in Fig. 1. In this neural model, the quantized input data are regarded as the quantum initial states. These initial states are inspired to be the quantum excited states using the mapping of the qubit neuron … we consider a n-dimensional real input vector x (k) = [xi(k), X2 (k) .. · Xn (k)]T … It can be converted into the quantum state … The quantum initial states, described as equation (2), are inspired into quantum transient sates using the rotating operation of the qubit neuron, then are inspired into the quantum excited states … These quantum excited states are connected with the quantized weights of the neurons in the competitive layer … the input-output relationship of the corresponding competitive-layer neurons can be described as
PNG
media_image1.png
52
286
media_image1.png
Greyscale
It is noteworthy that, yj(k) is not the actual output of the network, but the index value of the winning neuron in the competitive layer or the weight vector connected with the quantum excited state; It can be seen that the output y defining the inner product of the quantum state and the weights of the neurons of a lattice space of a neural network as shown in Fig. 1 and equation 9”], the estimated overlap values being derived from the measured outputs of the executed set of quantum circuits [page 3265, equation 9 shows the inner product overlap values is obtained from the output y
PNG
media_image1.png
52
286
media_image1.png
Greyscale
]; and
based on the estimated inner products measurement-derived estimated overlap values, identifying, by the system, a distance, of the neurons of the lattice space, to the sample quantum state within a Hilbert space of the neural network, the Hilbert space comprising the set of quantum data and the lattice space [Fig. 1, abstract, “A new quantum SOM (self-organizing map) neural model with elastic neighborhood radius is proposed. The new model objects the real input data into the quantum initial states. Through the operation of the quantum gate in the qubit neuron, the quantum initial states are converted into the quantum intermediate states, then into the quantum excited sates. These excited sates, being connected with the quantized weights, are perceived by the competitive-layer neurons. According to a new quantized competitive learning algorithm, the input data can be orderly topology mapped. The elastic neighborhood radius, in the new learning algorithm, is defined by both of the similarity and the distance between quantized weights and quantum excited states”; page 3264, Introduction, “The self-organizing map (SOM) [1,2] is an unsupervised neural network approach that reduces a high-dimensional data set to a representative and compact two-dimensional grid; pages 3264-3265, section 2.1, “The topology of this new neural model is shown in Fig. 1. In this neural model, the quantized input data are regarded as the quantum initial states. These initial states are inspired to be the quantum excited states using the mapping of the qubit neuron … we consider a n-dimensional real input vector x (k)”; Examiner interprets the high-dimensional or n-dimensional space as the Hilbert space which comprising quantum data and lattice space as shown in Fig. 1], wherein the quantum-state representations of the set of weight vectors reside within the Hilbert space [Fig. 1 shows the weight vectors associated with the quantum states within the Hilbert space].
Penghua does not explicitly teach
executing, a set of quantum circuits at a quantum processor;
based on the estimated inner products measurement-derived estimated overlap values, identifying, by the system, a best matching neuron of the lattice space having a closest distance, of the neurons of the lattice space, to the sample quantum state.
Corcoles-Gonzalez teaches
executes a set of quantum circuits at a quantum processor [Fig. 1, paragraph 0005, “a processor that executes computer executable components stored in a memory; a quantum feature map circuit component that estimates a kernel associated with a feature map”];
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include executing a set of quantum circuits at a quantum processor of Corcoles-Gonzalez. Doing so would help estimating a kernel associated with a feature map and performing a classification using the estimated kernel (Corcoles-Gonzalez, 0005).
Penghua and Corcoles-Gonzalez do not explicitly teach
based on the estimated inner products measurement-derived estimated overlap values, identifying, by the system, a best matching neuron of the lattice space having a closest distance, of the neurons of the lattice space, to the sample quantum state.
Kohonen teaches
based on the estimated inner products measurement- derived estimated overlap values, identifies a best matching neuron [page 1467, Col. 1, lines 3-7, “The simplest analytical measure for the match of x with the mi may be the inner product xmi”].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include identifying a best matching neuron based on the estimated inner products of Cohonen. Doing so would help obtaining the competitive learning rule of quantum SOM neural network (Penghua, 3265, section 2.2).
Penghua, Corcoles-Gonzalez and Kohonen do not explicitly teach
identifying, by the system, a best matching neuron of the lattice space having a closest distance, of the neurons of the lattice space, to the sample quantum state.
Asan teaches
identifies a best matching neuron of the lattice space having a closest distance, of the neurons of the lattice space, to the sample quantum state [page 300, last paragraph, “The output nodes (neurons) in a self-organizing map compete with each other to best represent the particular input sample. The success of representation is measured using a discriminant function, where an input vector is compared with the weight vector of each output node. The particular node with its connection weights most similar to the input sample is declared winner of the competition. There are a number of different functions to determine the winner, i.e. the best-matching unit (BMU) on the map. The most commonly used one is the Euclidean distance”; page 306, section 14.2.2.2 Similarity matching, “learning in SOM does not rely on predefined target outcomes that would guide the process. The output nodes compete among themselves to become activated. Only the node whose weight vector is most similar to the input vector will be activated and declared as the winner [20]. To find this best matching node, the distances between an input data (x) and all the weight vectors (w;) of the SOM are computed using different measurement methods, such as … Manhattan distance … Euclidean distance”].
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the quantum self-organizing map neural network model of Penghua to include identifying a best matching neuron of the lattice space having a closest distance of the neurons of the lattice space to the input vector of Asan. Doing so would help adjusting the weight vector of the best matching neuron and its neighborhood until the change of the weights is less than a threshold (Asan, pages 311-312, section 14.2.2.4).
Claim 16 is rejected by the same reason as of claim 11, since these claims recite the similar limitations.
Allowable Subject Matter
Claims 7 and 15 are objected to as being dependent upon the rejected base claims, but would be allowable if rewritten in independent form including all of the limitations of the base claims and any intervening claims.
The claims were previously indicated as having allowable subject matter, the examiner has performed an updated search, but no prior art references are found to teach or suggest, individually or in combination the limitations:
the estimating component executes additional sets of quantum circuits at the quantum processor resulting in additional measured outputs defining additional estimated inner products between additional sample quantum states, of the set of quantum data, and the set of weight vectors,
wherein the determining component, based on the additional estimated inner products, identifies additional best matching neurons to the additional sample quantum states, and
wherein the execution and the additional execution performed by the estimating component, and the identification and the additional identification performed by the determining component, together result in learning by the neural network of a mapping of the set of quantum data from a higher dimensional first continuous space to a lower dimensional second lattice space of neurons being a modification of the second output layer.
Therefore, the combination of features is considered to be allowable.
Prior Art
The prior art made of record and not relied upon is considered pertinent to applicant’s disclosure.
Greer (US Patent 8,041,651) describes a method of extracting spatial knowledge from an input signal.
Elfving et al. (US Pub. 2025/0181950) describes a method for solving regression problems using a quantum kernel.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to TRI T NGUYEN whose telephone number is 571-272-0103. The examiner can normally be reached M-F, 8 AM-5 PM, (CT).
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, OMAR FERNANDEZ can be reached at 571-272-2589. 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.
/TRI T NGUYEN/Examiner, Art Unit 2128
/OMAR F FERNANDEZ RIVAS/Supervisory Patent Examiner, Art Unit 2128