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 .
Priority
Applicant claims the benefit/priority from Korean Patent Application No.10-2023-0101075, filed on August 2, 2023, which is acknowledged.
Drawings
The drawings were received on 12/06/2023. These drawings are acceptable.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on the following date(s): 12/06/2023 has been considered by the examiner.
Response to Arguments
Applicant's arguments filed 7/09/2026 have been fully considered.
Regarding the rejection of claims under USC 35 102 and 103 the remarks are directed to amended limitations that have not been previously examined, see the rejection below that addresses the amended claim limitation.
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 1-2, 4-11 and 13-19 are rejected under 35 U.S.C. 103 as being unpatentable over Gunnels et al. (US 20200285947, hereinafter ‘Gu’) in view of Li at el. (NPL: An Image Classification Algorithm Based on Hybrid Quantum Classical Convolutional Neural Network, hereinafter ‘Li’).
Regarding independent claim 1, Gu teaches an apparatus with hybrid quantum-classical neural network architecture generation, the apparatus comprising: (in [0012] The illustrative embodiments provide a method, system, and computer program product for implementing a classical neural network with selective quantum computing kernel components. An embodiment of a method for implementing a hybrid classical-quantum neural network includes constructing, by at least a first processor, a neural network for classification of input data, the neural network including a plurality of neural network components… And in )
one or more hardware processors configured to generate a hybrid quantum-classical layer based neural network architecture based on setting information for generating a neural network architecture, (in [0012] The illustrative embodiments provide a method, system, and computer program product for implementing a classical neural network with selective quantum computing kernel components. An embodiment of a method for implementing a hybrid classical-quantum neural network includes constructing, by at least a first processor, a neural network for classification of input data, the neural network including a plurality of neural network components. The embodiment further includes initiating, by the at least a first processor, training of the neural network using training data. The embodiment further includes identifying, by the at least a first processor, one or more of the plurality of neural network components for replacement. The embodiment further includes constructing, by a quantum processor, a quantum component corresponding to the one or more network components. The embodiment still further includes replacing the one or more identified neural network components of the neural network with the quantum component to construct a hybrid classical-quantum neural network. Thus, the embodiment provides for implementing a classical neural network with selective quantum computing kernel components [one or more hardware processors configured to generate a hybrid quantum-classical layer based neural network architecture based on setting information for generating a neural network architecture] to improve classification of data using hybrid classical-quantum neural network. )
wherein, for the generating(in [0016] Another embodiment further includes monitoring a performance of the hybrid classical-quantum neural network to determine a quality level of classification results of the hybrid classical-quantum neural network. Another embodiment further includes identifying, responsive to determining that the quality level does not meet a threshold value, one or more other neural network components for replacement [wherein, for the generating of the hybrid quantum-classical layer based neural network architecture, the one or more hardware processors are configured to:, based on the setting information, whether to distribute neurons in each layer of the neural network into a quantum circuit and whether to distribute the neurons in the each layer into a classical circuit]. Thus, the embodiment provides for measuring a quality level of classification results produced by the hybrid classical-quantum neural network. And in [0036] In an embodiment, the classical neural network component is replaced with or duplicated by a quantum component such as a quantum kernel component that is equivalent to the classical neural network component [wherein, for the generating of the hybrid quantum-classical layer based neural network architecture, the one or more hardware processors are configured to:, based on the setting information, whether to distribute neurons in each layer of the neural network into a quantum circuit and whether to distribute the neurons in the each layer into a classical circuit]… And in [0038] In an embodiment, feature space extension of a neural network is achieved by modifying network components characterized to be indifferent to input of different classes. In the embodiment, various measures relying upon analysis of information flow and/or sensitivity are used for characterization/identification of classical subnetwork components [wherein, for the generating of the hybrid quantum-classical layer based neural network architecture, the one or more hardware processors are configured to:, based on the setting information, whether to distribute neurons in each layer of the neural network into a quantum circuit and whether to distribute the neurons in the each layer into a classical circuit] of the neural network whose feature space can benefit from a quantum feature space. Once a subnetwork component has been identified, the input and output of the of the subnetwork are “rewired” to a quantum kernelized feature space component implemented by a quantum processor In the embodiment, the quantum neural network involves parameters that are learnable by the full classical-quantum neural network structure [ wherein, for the generating]….)
verify a quantity the neurons of the each layer distributed into the quantum circuit and a quantity the neurons of the each layer distributed into the classical circuit, based on the determination of whether to distribute the neurons in the each layer into the quantum circuit or the classical circuit and on a result of distribution of the neurons in the each layer of the neural network into the quantum circuit and the classical circuit; generate a quantum circuit for the each layer, based on a the of distribution into the quantum circuit and on the verified quantities; and . (in [0038] In an embodiment, feature space extension of a neural network is achieved by modifying network components characterized to be indifferent to input of different classes. In the embodiment, various measures relying upon analysis of information flow and/or sensitivity are used for characterization/identification of classical subnetwork components [verify a quantity the neurons of the each layer distributed into the quantum circuit and a quantity the neurons of the each layer distributed into the classical circuit, based on the determination of whether to distribute the neurons in the each layer into the quantum circuit or the classical circuit and on a result of distribution of the neurons in the each layer of the neural network into the quantum circuit and the classical circuit] of the neural network whose feature space can benefit from a quantum feature space. Once a subnetwork component has been identified, the input and output of the of the subnetwork are “rewired” to a quantum kernelized feature space component implemented by a quantum processor In the embodiment, the quantum neural network involves parameters that are learnable by the full classical-quantum neural network structure [ verify a quantity the neurons of the each layer distributed into the quantum circuit and a quantity the neurons of the each layer distributed into the classical circuit, based on the determination of whether to distribute the neurons in the each layer into the quantum circuit or the classical circuit and on a result of distribution of the neurons in the each layer of the neural network into the quantum circuit and the classical circuit]….)
Gu teaches the process for generating an hybrid neural network processing architecture for distributing the structure of a neural network where it is determined that the network components and verified as claimed for the quantum structure component of the hybrid system by selecting the layers, comprising neurons to process on the quantum circuit, thus the remaining layers are determined to be part of the classical circuit.
Li teaches the component splitting where the fully connected layers are computed as layers of the classical circuit as selecting the layers, comprising neurons to process on the quantum circuit, thus the remaining layers are determined to be part of the classical circuit. depicted in Fig. 5:
PNG
media_image1.png
528
1000
media_image1.png
Greyscale
Figure 5: Hybrid quantum-classical convolutional neural network architecture [verify a quantity the neurons of the each layer distributed into the quantum circuit and a quantity the neurons of the each layer distributed into the classical circuit, based on the determination of whether to distribute the neurons in the each layer into the quantum circuit or the classical circuit and on a result of distribution of the neurons in the each layer of the neural network into the quantum circuit and the classical circuit; generate a quantum circuit for the each layer, based on a the of distribution into the quantum circuit and on the verified quantities; and ].
And in Sec. 3.4. Generating Hybrid Network. HQCCNN consists of a quantum convolutional layer, a quantum pooling layer, and a classical fully connected layer [verify a quantity the neurons of the each layer distributed into the quantum circuit and a quantity the neurons of the each layer distributed into the classical circuit, based on the determination of whether to distribute the neurons in the each layer into the quantum circuit or the classical circuit and on a result of distribution of the neurons in the each layer of the neural network into the quantum circuit and the classical circuit]. As shown in Figure 5 [generate a quantum circuit for the each layer, based on a the of distribution into the quantum circuit and on the verified quantities; and ], the quantum convolutional layer consists of multiple convolution kernels, which complete the quantum convolution to obtain the feature map. The convolution kernel in Figure 5 is the convolution structure introduced in Section 3.2, but the parameters of different convolution kernels are different. The convolution results are reduced by the quantum pooling layer, and the quantum pooling unit is the quantum pooling structure introduced in Section 3.3. Then measuring the specific qubits, the measurement results are input into the fully connected layer to obtain the image’s class.
Li and Gu are analogous art because both involve developing information retrieval and object recognition techniques using machine learning systems and algorithms.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of the prior art of developing information processing techniques/systems for image recognition method and device based on quantum classical hybrid neural network, as disclosed by Li with the method of developing information processing techniques/systems method for implementing a hybrid classical-quantum neural network, as disclosed by Gu.
One of ordinary skill in the arts would have been motivated to combine the methods disclosed by Li and Gu , as noted above. Doing so allows for constructing quantum machine learning algorithms, that achieve better algorithm performance than classical algorithms, (Li, Abstract).
Regarding claim 2, the rejection of claim 1 is incorporated and Gu in combination with Li further teaches the apparatus of claim 1, wherein the setting information for generating the architecture comprises one or more of a total number of layers of a neural network architecture to be generated, a number of input neurons and output neurons in each layer, a probability that the respective neurons in the each layer are distributed into the quantum circuit or the classical circuit, and a number of output neurons of each of the quantum circuit and the classical circuit in the each layer. (in [0072] With reference to FIGS. 3A-3C, these figures depict s simplified example sequence for replacing components of a classical components of a neural network 300 with quantum computing kernel components in accordance with an illustrative embodiment. FIG. 3A illustrates an initial state of a neural network 300 formed of a classical neural network having a number of nodes interconnected in a plurality of layers including an input layer [wherein the setting information for generating the architecture comprises one or more of a total number of layers of a neural network architecture to be generated, a number of input neurons and output neurons in each layer], an output layer, and a number of hidden layers of nodes between the input layer and the output layer [wherein the setting information for generating the architecture comprises one or more of a total number of layers of a neural network architecture to be generated]. In one or more embodiments, the classical neural network is implemented by a classical computer such as classical processing system 104.)
Regarding claim 4, the rejection of claim 1 is incorporated and Gu in combination with Li further teaches the apparatus of claim 1, wherein the one or more hardware processors are configured to measure performance of the generated neural network architecture. (in [0096] In block 916, classical processor 122 monitors the performance of the joint classical-quantum neural network to determine a quality level [wherein the one or more hardware processors are configured to measure performance of the generated neural network architecture] of the classification results of the joint classical-quantum neural network. In block 918, classical processor 122 determines whether the quality level of the joint classical-quantum neural network is at an acceptable threshold value.)
Regarding claim 5, the rejection of claim 4 is incorporated and Gu in combination with Li further teaches the apparatus of claim 4, wherein the one or more hardware processors are configured to measure performance of the classical circuit; and measure performance of the quantum circuit. (in [0012] The illustrative embodiments provide a method, system, and computer program product for implementing a classical neural network with selective quantum computing kernel components. An embodiment of a method for implementing a hybrid classical-quantum neural network [wherein the one or more hardware processors are configured to measure performance of the classical circuit; and measure performance of the quantum circuit] includes constructing, by at least a first processor, a neural network for classification of input data, the neural network including a plurality of neural network components. The embodiment further includes initiating, by the at least a first processor, training of the neural network using training data. The embodiment further includes identifying, by the at least a first processor, one or more of the plurality of neural network components for replacement [measure performance of the classical circuit as performance for intensifying components to replace]. The embodiment further includes constructing, by a quantum processor, a quantum component corresponding to the one or more network components [measure performance of the quantum circuit as the quantum components performance corresponding to the components for replacement]. The embodiment still further includes replacing the one or more identified neural network components of the neural network with the quantum component to construct a hybrid classical-quantum neural network. Thus, the embodiment provides for implementing a classical neural network with selective quantum computing kernel components to improve classification of data using hybrid classical-quantum neural network.)
Regarding claim 6, the rejection of claim 5 is incorporated and Gu in combination with Li further teaches the apparatus of claim 5, wherein the one or more hardware processors are configured to control quantum states of quantum objects in the QPU, by using quantum technology. (in [0005] A quantum processor (q-processor) uses the unique nature of entangled qubit devices [wherein the one or more hardware processors are configured to control quantum states of quantum objects in the QPU, by using quantum technology] (compactly referred to herein as “qubit,” plural “qubits”) to perform computational tasks. In the particular realms where quantum mechanics operates, particles of matter can exist simultaneously in multiple states-such as an “on” state, an “off” state, and both “on” and “off” states simultaneously. Where binary computing using semiconductor processors is limited to using just the on and off states (equivalent to 1 and 0 in binary code), a quantum processor harnesses these quantum states of matter [further comprising a quantum controller configured to control quantum states of quantum objects in the QPU, by using quantum technology] to output signals that are usable in data computing.)
Regarding claim 7, the rejection of claim 4 is incorporated and Gu in combination with Li further teaches the apparatus of claim 4, wherein the one or more hardware processors are configured to determine performance information of the generated neural network architecture by using the computing device, and to determine whether performance of the generated neural network architecture satisfies a target performance by using the determined performance information. (in [0096] In block 916, classical processor 122 monitors the performance of the joint classical-quantum neural network to determine a quality level of the classification results of the joint classical-quantum neural network. In block 918, classical processor 122 determines whether the quality level [wherein the one or more hardware processors are configured to determine performance information of the generated neural network architecture by using the computing device] of the joint classical-quantum neural network is at an acceptable threshold value [and to determine whether performance of the generated neural network architecture satisfies a target performance by using the determined performance information].)
Regarding claim 8, the rejection of claim 7 is incorporated and Gu in combination with Li further teaches the apparatus of claim 7, wherein the performance information comprises one or more indicators comprising any one or any combination of any two or more of learnability of the generated neural network architecture, training validation accuracy, training time, inference accuracy, and inference time of the generated neural network architecture. (in [0076] In the illustrated embodiment, solution 412 is evaluated to determine if classical neural network with quantum component 410 provides results of acceptable quality. If an acceptable quality has not been achieved, training operations 402 evolve parameters of the neural network which may include further substituting a classical component of the neural network with a quantum component and/or substituting a quantum component with a classical component in order to improve accuracy and/or efficiency [wherein the performance information comprises one or more indicators comprising any one or any combination of any two or more of learnability of the generated neural network architecture, … inference accuracy]. Training operations 402 then continue until acceptable results are achieved. Once acceptable results are achieved, a trained neural network with quantum components is output.)
Regarding claim 9, the rejection of claim 7 is incorporated and Gu in combination with Li further teaches the apparatus of claim 7, wherein in response to the target performance being satisfied, the one or more hardware processors are is configured to determine the generated neural network architecture to be a final neural network architecture. (in [0076] In the illustrated embodiment, solution 412 is evaluated to determine if classical neural network with quantum component 410 provides results of acceptable quality. If an acceptable quality has not been achieved, training operations 402 evolve parameters of the neural network which may include further substituting a classical component of the neural network with a quantum component and/or substituting a quantum component with a classical component in order to improve accuracy and/or efficiency [wherein in response to the target performance being satisfied, the performance comparison module is configured to determine the generated neural network architecture to be a final neural network architecture]. Training operations 402 then continue until acceptable results are achieved. Once acceptable results are achieved, a trained neural network with quantum components is output [wherein in response to the target performance being satisfied, the performance comparison module is configured to determine the generated neural network architecture to be a final neural network architecture as the final output network network].)
Regarding claim 10, the rejection of claim 7 is incorporated and Gu in combination with Li further teaches the apparatus of claim 7, wherein in response to the target performance not being satisfied, the one or more hardware processors are configured to store the setting information, result information of generating the neural network architecture, and the performance information in a memory. (in [0076] In the illustrated embodiment, solution 412 is evaluated to determine if classical neural network with quantum component 410 provides results of acceptable quality. If an acceptable quality has not been achieved [wherein in response to the target performance not being satisfied,..], training operations 402 evolve parameters of the neural network which may include further substituting a classical component of the neural network with a quantum component and/or substituting a quantum component with a classical component in order to improve accuracy and/or efficiency [the performance comparison module is configured to store the setting information, result information of generating the neural network architecture, and the performance information in a memory as stored information for repeating the training operations as a next iteration until the target performance is satisfied]. Training operations 402 then continue until acceptable results are achieved. Once acceptable results are achieved, a trained neural network with quantum components is output.)
Regarding claim 11, the rejection of claim 7 is incorporated and Gu in combination with Li further teaches the apparatus of claim 7, wherein the one or more hardware processors are configured to, in response to the target performance not being satisfied: update the setting information by using new setting information; and generate the neural network architecture again. (in [0076] In the illustrated embodiment, solution 412 is evaluated to determine if classical neural network with quantum component 410 provides results of acceptable quality. If an acceptable quality has not been achieved [wherein the one or more hardware processors are configured to, in response to the target performance not being satisfied:,..], training operations 402 evolve parameters of the neural network which may include further substituting a classical component of the neural network with a quantum component and/or substituting a quantum component with a classical component in order to improve accuracy and/or efficiency [update the setting information by using new setting information; and generate the neural network architecture again as new distributed components of the neural network updated when the target performance is not satisfied]. Training operations 402 then continue until acceptable results are achieved. Once acceptable results are achieved, a trained neural network with quantum components is output.)
Regarding independent claim 13, Gu teaches a processor-implemented method with hybrid quantum-classical neural network architecture generation, the method comprising: (in [0012] The illustrative embodiments provide a method, system, and computer program product for implementing a classical neural network with selective quantum computing kernel components. An embodiment of a method for implementing a hybrid classical-quantum neural network includes constructing, by at least a first processor, a neural network for classification of input data, the neural network including a plurality of neural network components…)
receiving setting information for generating a neural network architecture; and generating a hybrid quantum-classical layer based neural network architecture based on the setting information, (in [0012] The illustrative embodiments provide a method, system, and computer program product for implementing a classical neural network with selective quantum computing kernel components. An embodiment of a method for implementing a hybrid classical-quantum neural network includes constructing, by at least a first processor, a neural network for classification of input data, the neural network including a plurality of neural network components. The embodiment further includes initiating, by the at least a first processor, training of the neural network using training data. The embodiment further includes identifying [receiving setting information for generating a neural network architecture; and generating a hybrid quantum-classical layer based neural network architecture based on the setting information], by the at least a first processor, one or more of the plurality of neural network components for replacement. The embodiment further includes constructing, by a quantum processor, a quantum component corresponding to the one or more network components. The embodiment still further includes replacing the one or more identified neural network components of the neural network with the quantum component to construct a hybrid classical-quantum neural network. Thus, the embodiment provides for implementing a classical neural network with selective quantum computing kernel components [receiving setting information for generating a neural network architecture; and generating a hybrid quantum-classical layer based neural network architecture based on the setting information] to improve classification of data using hybrid classical-quantum neural network. )
wherein the generating of the neural network architecture comprises: determining, based on the setting information, whether to distribute neurons in each layer of the neural network into a quantum circuit and whether to distribute the neurons in the each layer into a classical circuit; (in [0016] Another embodiment further includes monitoring a performance of the hybrid classical-quantum neural network to determine a quality level of classification results of the hybrid classical-quantum neural network. Another embodiment further includes identifying, responsive to determining that the quality level does not meet a threshold value, one or more other neural network components for replacement [wherein the generating of the neural network architecture comprises: determining, based on the setting information, whether to distribute neurons in each layer of the neural network into a quantum circuit and whether to distribute the neurons in the each layer into a classical circuit]. Thus, the embodiment provides for measuring a quality level of classification results produced by the hybrid classical-quantum neural network. And in [0036] In an embodiment, the classical neural network component is replaced with or duplicated by a quantum component such as a quantum kernel component that is equivalent to the classical neural network component [wherein the generating of the neural network architecture comprises: determining, based on the setting information, whether to distribute neurons in each layer of the neural network into a quantum circuit and whether to distribute the neurons in the each layer into a classical circuit]…)
The remaining claim 13 limitations are similar to those in claim 1 and rejected under the same rationale.
Regarding claims 14, the limitations are similar to those in claim 2, and are thus rejected under the same rationale.
Regarding claims 15, the limitations are similar to those in claim 3, and are thus rejected under the same rationale.
Regarding claims 16, the limitations are similar to those in claim 7, and are thus rejected under the same rationale.
Regarding claims 17, the limitations are similar to those in claim 10, and are thus rejected under the same rationale.
Regarding claims 18, the limitations are similar to those in claim 11, and are thus rejected under the same rationale.
Regarding claims 19, the limitations are similar to those in claim 12, and are thus rejected under the same rationale.
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over Gu in view of Li in further view of Baughman et al. (US 20230065684, hereinafter ‘Ba’).
Regarding claim 12, the rejection of claim 11 is incorporated and Gu in combination with Li further teaches the apparatus of claim 11, wherein in response to the target performance not being satisfied, the one or more hardware processors are configured to: determine whether a number of times that neural network architectures are generated so far exceeds a predetermined number of times; and in response to the predetermined number of times being exceeded, determine one of the neural network architectures, generated so far, to be a final neural network architecture based on the performance information. (in [0076] In the illustrated embodiment, solution 412 is evaluated to determine if classical neural network with quantum component 410 provides results of acceptable quality. If an acceptable quality has not been achieved [wherein in response to the target performance not being satisfied, the performance comparison module is configured to: determine whether a number of times that neural network architectures are generated so far exceeds a predetermined number of times determined to exceed one iteration as a predetermined number of time for making a determination of the quality level so far], training operations 402 evolve parameters of the neural network which may include further substituting a classical component of the neural network with a quantum component and/or substituting a quantum component with a classical component in order to improve accuracy and/or efficiency [determine whether a number of times that neural network architectures are generated so far exceeds a predetermined number of times and in response to the predetermined number of times being exceeded, determine one of the neural network architectures, generated so far, to be a final neural network architecture based on the performance information as new distributed components of the neural network as a training iteration]]. Training operations 402 then continue until acceptable results are achieved. Once acceptable results are achieved, a trained neural network with quantum components is output.)
One of ordinary skill in the art would understand that training iterations can be terminated based on exceeding the first iteration by a predetermined amount.
Additionally Ba teaches determine whether a number of times that neural network architectures are generated so far exceeds a predetermined number of times and in response to the predetermined number of times being exceeded, determine one of the neural network architectures, in [0037] An embodiment uses one or more of the feature groups and a quantum data model executing on a quantum processor to select and score another set of data attributes, or features, from input training data. One embodiment scores a group of features using a quantum deep neural network (QDNN) [determine one of the neural network architectures], a deep neural network implemented with quantum neural network layers (neural network layers implemented using a quantum processor) instead of a classical deep neural network's layers implemented using a classical processor. The output of the QDNN is fed into a classical deep neural network for further processing… Once one or more groups of features have been scored using a quantum processor, an embodiment merges the quantum processor-scored groups of features into a combined set of scored quantum features. [0038] An embodiment adjusts one or more scores of quantum features according to an accuracy of the quantum data model. A quantum data model executing on a quantum processor produces probabilistic results. In other words, if the model is executed multiple times with the same set of parameter values and initial conditions, model output will not be exactly the same after each execution. Instead, the set of model outputs forms a probability distribution. Thus, an embodiment executes a quantum data model on training data for a number of iterations [determine whether a number of times that neural network architectures are generated so far exceeds a predetermined number of times and in response to the predetermined number of times being exceeded], combines the results of the iterations using a presently known technique (e.g., by averaging them), and compares the results to known results of feature scores using the training data to determine the model's accuracy with respect to a particular feature [determine one of the neural network architectures, generated so far, to be a final neural network architecture based on the performance information associated with the predetermined iterations far exceeding first predetermined iteration], denoted as accuracy.sub.ki. Another embodiment determines the model's accuracy with respect to a particular feature using another presently known technique
Ba, Li and Gu are analogous art because both involve developing information retrieval and processing techniques/systems using a hybrid classical-quantum computing environment.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of the prior art of developing information processing techniques/systems for combined classical/quantum predictor evaluation with model accuracy adjustment, as disclosed by Ba with the method of developing information processing techniques/systems method for implementing a hybrid classical-quantum neural network, as collectively disclosed by Li and Gu.
One of ordinary skill in the arts would have been motivated to combine the methods disclosed by Ba, Li and Gu , as noted above. Doing so allows for enabling predictor valuation of a classical/quantum predictor that is more accurate and more quickly computable than the techniques currently available., (Ba, 0029).
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Gu in view of Li in further view of Dou et al. (US 20240095563, hereinafter ‘Dou’).
Regarding independent claim 20, Gu teaches an electronic device comprising: including a hybrid quantum-classical neural network architecture generated by an apparatus for generating a hybrid quantum-classical neural network architecture; (in [0016] Another embodiment further includes monitoring a performance of the hybrid classical-quantum neural network to determine a quality level of classification results of the hybrid classical-quantum neural network. Another embodiment further includes identifying, responsive to determining that the quality level does not meet a threshold value, one or more other neural network components for replacement. Thus, the embodiment provides for measuring a quality level of classification results produced by the hybrid classical-quantum neural network…)
and one or more hardware processors configured to operate the hybrid quantum-classical neural network device, (in [0012] The illustrative embodiments provide a method, system, and computer program product for implementing a classical neural network with selective quantum computing kernel components. An embodiment of a method for implementing a hybrid classical-quantum neural network includes constructing, by at least a first processor, a neural network for classification of input data, the neural network including a plurality of neural network components. The embodiment further includes initiating, by the at least a first processor, training of the neural network using training data. The embodiment further includes identifying, by the at least a first processor, one or more of the plurality of neural network components for replacement. The embodiment further includes constructing, by a quantum processor, a quantum component corresponding to the one or more network components.… [0018] In another embodiment, the at least a first processor comprises a classical processor. In another embodiment, the neural network comprises a classical neural network… )
and to perform any one or any combination of any two or more of image processing, natural language processing, and applying artificial intelligence and machine learning, (in [0034] A deep neural network (DNN) is an artificial neural network (ANN) with multiple hidden layers of units between the input and output layers. Similar to shallow ANNs, DNNs can model complex non-linear relationships. DNN architectures, e.g., for object detection and parsing, generate compositional models where the object is expressed as a layered composition of image primitives... DNNs are often used for image classification tasks for computer vision in which an object represented in an image is identified and classified [to perform any one or any combination of any two or more of image processing, ]… )
wherein the hybrid quantum-classical neural network is generated by stochastically distributing neurons in one or more layers of the neural network into both a quantum circuit and a classical circuit. (in [0091] … In an embodiment, analysis of information flow and/or sensitivity are used for characterization/identification of classical network components of the neural network whose feature space can benefit from a quantum feature space enhancement [wherein the hybrid quantum-classical neural network is generated by stochastically distributing neurons in one or more layers of the neural network into both a quantum circuit and a classical circuit]. In particular embodiments, classical processor 122 may determine that classical network components that are relatively insensitive to varying input signals are candidates for replacement by a quantum kernel component. [0092] In particular embodiments, terminal classical components of the classical neural network are analyzed differently than non-terminal components. In a particular embodiment, for terminal nodes/components of the neural network [stochastically distributing neurons in one or more layers of the neural network into both a quantum circuit and a classical circuit], classical processor 122 monitors the response (e.g., changes in which class is chosen) of a classical neural network node to “noise” such as random noise [wherein the hybrid quantum-classical neural network is generated by stochastically distributing neurons in one or more layers of the neural network into both a quantum circuit and a classical circuit] or actual samples that are close to one another using a particular vector-based metric of distance…; And distribution based on a gradient stochastic learning process for selection hybrid components in [0094] In view of this, classical processor 122 may determine by tracking the propagation of the signal from input or from traversing backwards in the neural network level-by-level from the component at depth D) that the node(s) that are tamping out the sensitivity and look at the nodes as a new candidate(s) for replacement. If the component at depth D is receiving a different signal, but the difference does not exceed a particular threshold, noise may be added to the signals incoming to the node. In other particular embodiments, classical processor 112 may monitor the backpropagation [wherein the hybrid quantum-classical neural network is generated by stochastically distributing neurons in one or more layers of the neural network into both a quantum circuit and a classical circuit] in the neural network to determine that a component is insensitive to feedback from training.)
The remaining limitations are similar to the claim 1 limitations examined above and rejected are under the same rationale.
While one of ordinary skill in the art would interpret the hybrid quantum-classical neural network for object recognition is also be a machine learning process.
Gu does not expressly disclose the use of the hybrid quantum-classical neural network for object recognition as a machine learning process.
Dou does expressly teach the use of the hybrid quantum-classical neural network for object recognition as a machine learning process, in [0154] inputting the target image data into a pre-trained quantum classical hybrid neural network for image recognition [to perform any one or any combination of any two or more of image processing, intelligence and machine learning], wherein the quantum classical hybrid neural network comprises: a feature learning module and a classifying module, and the feature learning module [to perform any one or any combination of any two or more of image processing, ]…
Additionally, Dou also teaches natural language processing, in [0256] Specifically, the target object may include, but is not limited to, an image, natural language, or audio information, such as an image of each frame of a video, an individually shot picture, a language text of a chat record, voice information, and so on. Attribute information refers to information that carries and embodies the implicit characteristics of a target object. Taking an image as an example, the attribute information of the image includes pixel information, contour information, color information, label information, etc., and the attribute value is the visualized data value corresponding to the attribute information, such as pixel value, label value, etc.
Dou, Li and Gu are analogous art because both involve developing information retrieval and processing techniques/systems using a hybrid classical-quantum computing environment.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of the prior art of developing information processing techniques/systems for image recognition method and device based on quantum classical hybrid neural network, as disclosed by Dou with the method of developing information processing techniques/systems method for implementing a hybrid classical-quantum neural network, as collectively disclosed by Li and Gu.
One of ordinary skill in the arts would have been motivated to combine the methods disclosed by Dou, Li and Gu , as noted above. Doing so enables application of quantum computing in the field of convolution neural network models, so as to take advantage of the parallelism of quantum computing and for enabling image processing learning task, (Dou, 0010 & 0089).
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.
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Zhao et al. (NPL From IDS filed 12/06/2023, “QDNN: deep Neural networks with quantum layers”): teaches in pg. 4 of 9: Sec: 2.2: The structure of quantum neural network layers should not be randomly chosen because there exist a barren plateaus, which makes the model untrainable.
Aspuru-Guzik et al. (US 20200410384): teaches Hybrid quantum-classical generative models for learning data distributions are provided.
Arrazola et al. (US 20240169235): teaches a hybrid quantum-classical computing method and system that leverages quantum processing to generate data that enables the training of neural networks for the purpose of density functional theory (DFT) functional determination. Physical systems are modeled on a quantum processing module and simulated to generate energy values and electronic density functions as training data with sufficient degrees of quality and accuracy. The training data may be in classical form and are used to train a neural network. The trained neural network may then be employed to parameterize DFT functionals.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to OLUWATOSIN ALABI whose telephone number is (571)272-0516. The examiner can normally be reached Monday-Friday, 8:00am-5:00pm EST..
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, Michael Huntley can be reached at (303) 297-4307. 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.
/OLUWATOSIN ALABI/ Primary Examiner, Art Unit 2129
Prosecution Insights