DETAILED ACTION
This action is in response to the original filling on 05/23/2023. Claims 1-19 are pending for examination.
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 .
Claim Objections
Claims 2 and 12 objected to because of the following informalities: wherein the training dataset comprises items, item labels, andother meta-information corresponding to each item. The words “andother” are improperly spaced. Appropriate correction is required.
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 9-10 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 applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding Claim 9, Claim 9 recites: when checking the augmented training dataset for sufficiency to train the NN, and determining that the augmented training dataset is insufficient to train the NN, executing the method of claim 1. It is not clearly established what executing Claim 1 means in relation to checking and determining. It is unclear if both conditions are required for executing Claim 1, or if Claim 1 can be executed before these conditions occur. For the purposes of examination, “the” will be placed before “executing” in order to create an order of operations wherein execution of Claim 1 occurs after the checking and determination steps.
Regarding Claim 10, Claim 10 recites: when checking the augmented training dataset for sufficiency to train the NN, and determining that the augmented training dataset is sufficient to train the NN, training the NN. It is not clearly established what training the neural network means in relation to checking and determining. It is unclear if both conditions are required for training the neural network, or if training can be executed before these conditions occur. For the purposes of examination, “the” will be placed after “training” in order to create an order of operations wherein training occurs after the checking and determination steps.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are: a data storage configured to, GCNN configured to, augmenter configured to, in Claim 11, sufficiency checker configured to, in Claim 13, and sufficiency checker is further configured to in Claim 18.
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
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 1-19 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim 1
Step 1: Claim 1 recite a method, and so is directed to the statutory categories of a method.
Step 2A Prong 1: The claim recites, inter alia: “using at least one sufficiency criteria, determining that the training dataset is insufficient for training the neural network NN; selecting a generative convolutional neural network GCNN for which the existing training dataset is sufficient to get trained; generating, an additional item; and adding the additional item to the training dataset.” Under its broadest reasonable interpretation, these limitations encompass the mental process of determination or judgement that is practically capable of being performed in the human mind with the aid of pen and paper.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. The additional element of: “gaining access to a training dataset for training a neural network NN;” amounts to no more than mere data gathering (see MPEP § 2106.05(g)). The claim invokes computers or other machinery merely as a tool to perform an existing process. The additional element of: “training the generative convolutional neural network GCNN using the training dataset or a subset of the training dataset;” amounts to no more than a high level of generality “Apply it” or an equivalent (see MPEP § 2106.05(f)). The additional element of: “by the GCNN trained on the existing training dataset,” amounts to no more than mere instructions to apply an exception, particularly using a generic computer algorithm (see MPEP § 2106.05(g)).
Step 2B: The claim does not contain significantly more than the judicial exception. The additional element of: “gaining access to a training dataset for training a neural network NN;” amounts to no more than mere data gathering (see MPEP § 2106.05(g)). The claim invokes computers or other machinery merely as a tool to perform an existing process. The additional element of: “training the generative convolutional neural network GCNN using the training dataset or a subset of the training dataset;” amounts to no more than a high level of generality “Apply it” or an equivalent (see MPEP § 2106.05(f)). The additional element of: “by the GCNN trained on the existing training dataset,” amounts to no more than mere instructions to apply an exception, particularly using a generic computer algorithm (see MPEP § 2106.05(f)). Nothing in the claims provides significantly more than the abstract idea. As such, the claim is ineligible.
Claim 11
Step 1: Claim 11 recites a system, and so is directed to the statutory categories of a system.
Step 2A Prong 1: The claim recites, inter alia: “add the new generated item to the training dataset.” Under its broadest reasonable interpretation, these limitations encompass the mental process of determination or judgement that is practically capable of being performed in the human mind with the aid of pen and paper.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. The additional element of: “a data storage configured to store a training dataset;” amounts to no more than mere instructions to apply an exception, particularly using a generic computer component (see MPEP § 2106.05(f)). The additional element of: “a data generator comprising a generative convolutional neural network GCNN configured to be trained on a same type of datasets as the NN and to output, after training, a new generated item; and” amounts to no more than a high level of generality “Apply it” or an equivalent (see MPEP § 2106.05(f)). Whether the claim recites only the idea of a solution or outcome i.e., the claim fails to recite details of how a solution to a problem is accomplished. The additional element of: “a dataset augmenter configured to” amounts to no more than mere instructions to apply an exception, particularly using a generic computer component (see MPEP § 2106.05(f)).
Step 2B: The claim does not contain significantly more than the judicial exception. The additional element of: “a data storage configured to store a training dataset;” amounts to no more than mere instructions to apply an exception, particularly using a generic computer component (see MPEP § 2106.05(f)). The additional element of: “a data generator comprising a generative convolutional neural network GCNN configured to be trained on a same type of datasets as the NN and to output, after training, a new generated item; and” amounts to no more than a high level of generality “Apply it” or an equivalent (see MPEP § 2106.05(f)). Whether the claim recites only the idea of a solution or outcome i.e., the claim fails to recite details of how a solution to a problem is accomplished. The additional element of: “a dataset augmenter configured to” amounts to no more than mere instructions to apply an exception, particularly using a generic computer component (see MPEP § 2106.05(f)). As such, the claim is ineligible.
Claims 2 and 12
Step 1: Claims recite a method and a system, and so are directed to the statutory categories of a method and a system.
Step 2A Prong 1: Claims 2 and 12 merely narrow the previously recited abstract limitations. For the reasons described above with respect to claims 1 and 11, this judicial exception is not meaningfully integrated into a practical application, or significantly more than the abstract idea. The claims disclose similar limitations described for the independent claims above and do not provide anything more than the mental processes that are practically capable of being performed in the human mind with the assistance of pen and paper and mathematical concepts that are achievable through mathematical computation.
Step 2A Prong 2: Claims 2 and 12 recite the additional element of: “wherein the training dataset comprises items, item labels, andother [sic] meta-information corresponding to each of the items.” amounts to no more than insignificant extra solution activity, particularly selecting a particular data source (see MPEP § 2106.05(g)).
Step 2B: The claims do not contain significantly more than the judicial exception.
Claims 3 and 13
Step 1: Claims recite a method and a system, and so are directed to the statutory categories of a method and a system.
Step 2A Prong 1: The claims recite, inter alia: “wherein the gaining access to the insufficient training dataset for training a neural network NN further comprises identifying if the training dataset is sufficient for training the neural network NN based on a sufficiency criterion.” Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determination or judgement that is practically capable of being performed in the human mind with the aid of pen and paper.
Step 2A Prong 2 and Step 2B: There are no additional elements recited so the claim does not provide a practical application and is not considered to be significantly more. As such, the claims are patent ineligible.
Claims 4 and 14
Step 1: Claims recite a method and a system, and so are directed to the statutory categories of a method and a system.
Step 2A Prong 1: Claims 4 and 14 merely narrow the previously recited abstract limitations. For the reasons described above with respect to claims 1 and 11, this judicial exception is not meaningfully integrated into a practical application, or significantly more than the abstract idea. The claims disclose similar limitations described for the independent claims above and do not provide anything more than the mental processes that are practically capable of being performed in the human mind with the assistance of pen and paper and mathematical concepts that are achievable through mathematical computation.
Step 2A Prong 2: Claims 4 and 14 recite the additional element of: “wherein the GCNN has the same degree of freedom as NN.” amounts to no more than insignificant extra solution activity, particularly selecting a particular data source (see MPEP § 2106.05(g)).
Step 2B: The claims do not contain significantly more than the judicial exception.
Claims 5 and 15
Step 1: Claims recite a method and a system, and so are directed to the statutory categories of a method and a system.
Step 2A Prong 1: The claims recite, inter alia: “wherein the degrees of freedom of the GCNN is decreased from the degree of freedom of NN to require fewer items in a GCNN training set to make the existing training set sufficient for the GCNN training.” Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of mathematical determination that is practically capable of being performed in the human mind or with the aid of pen and paper.
Step 2A Prong 2 and Step 2B: There are no additional elements recited so the claim does not provide a practical application and is not considered to be significantly more. As such, the claims are patent ineligible.
Claims 6 and 16
Step 1: Claims recite a method and a system, and so are directed to the statutory categories of a method and a system.
Step 2A Prong 1: The claims recite, inter alia: “wherein different parameters comprising degrees of freedom are omitted in the GCNN to generate consecutive new items.” Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of mathematical determination that is practically capable of being performed in the human mind or with the aid of pen and paper.
Step 2A Prong 2 and Step 2B: There are no additional elements recited so the claim does not provide a practical application and is not considered to be significantly more. As such, the claims are patent ineligible.
Claims 7 and 17
Step 1: Claims recite a method and a system, and so are directed to the statutory categories of a method and a system.
Step 2A Prong 1: The claims recite, inter alia: “wherein the omitted degrees of freedom in the GCNN are determined based on statistical parameters related to different parameters of items within the training set.” Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of mathematical determination that is practically capable of being performed in the human mind or with the aid of pen and paper.
Step 2A Prong 2 and Step 2B: There are no additional elements recited so the claim does not provide a practical application and is not considered to be significantly more. As such, the claims are patent ineligible.
Claims 8 and 18
Step 1: Claims recite a method and a system, and so are directed to the statutory categories of a method and a system.
Step 2A Prong 1: The claims recite, inter alia: “checking the augmented training dataset for sufficiency to train the NN after adding the at least one additional item to the original training set.” Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of evaluation or judgement that is practically capable of being performed in the human mind or with the aid of pen and paper.
Step 2A Prong 2 and Step 2B: There are no additional elements recited so the claim does not provide a practical application and is not considered to be significantly more. As such, the claims are patent ineligible.
Claims 9 and 19
Step 1: Claims recite a method and a system, and so are directed to the statutory categories of a method and a system.
Step 2A Prong 1: Claims 9 and 19 merely narrow the previously recited abstract limitations. For the reasons described above with respect to claims 8 and 18, this judicial exception is not meaningfully integrated into a practical application, or significantly more than the abstract idea. The claims disclose similar limitations described for the independent claims above and do not provide anything more than the mental processes that are practically capable of being performed in the human mind with the assistance of pen and paper and mathematical concepts that are achievable through mathematical computation.
Step 2A Prong 2: Claims 9 and 19 recite the additional element of: “when checking the augmented training dataset for sufficiency to train the NN, and determining that the augmented training dataset is insufficient to train the NN, executing the method of claim 1.” amounts to no more than linking an exception to a particular technological environment (see MPEP § 2106.05(h)).
Step 2B: The claims do not contain significantly more than the judicial exception.
Claim 10
Step 1: Claim recites a method, and so is directed to the statutory category of a method.
Step 2A Prong 1: Claim 10 merely narrows the previously recited abstract limitations. For the reasons described above with respect to claims 8, this judicial exception is not meaningfully integrated into a practical application, or significantly more than the abstract idea. The claims disclose similar limitations described for the independent claims above and do not provide anything more than the mental processes that are practically capable of being performed in the human mind with the assistance of pen and paper and mathematical concepts that are achievable through mathematical computation.
Step 2A Prong 2: Claim 8 recites, inter alia: “when checking the augmented training dataset for sufficiency to train the NN, and determining that the augmented training dataset is sufficient to train the NN, training the NN.” amounts to no more than linking an exception to a particular technological environment (see MPEP § 2106.05(h)).
Step 2B: The claims do not contain significantly more than the judicial exception.
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 1-19 are rejected under 35 U.S.C. 103 as being unpatentable over Cella et al. (US 20250054008 A1 hereinafter Cella) in view of Bradley et al. (US 12488524 B2 hereinafter Bradley).
Regarding Claim 1, Bradley teaches a method for automatically augmenting a machine training dataset, (Cella Paragraph [0525] In some embodiments, the machine learning model 3000 may automatically increase or decrease collection rates, processing, storage, sampling rates, bandwidth allocation, bitrates, and other attributes of sensor data collection to achieve or better achieve the modeling goal. Paragraph [0928] In embodiments, the reconfiguring/retraining an executive agent may include removing an input that is the source of the error, reconfiguring a set of nodes of the artificial intelligence system, reconfiguring a set of weights of the artificial intelligence system, reconfiguring a set of outputs of the artificial intelligence system, reconfiguring a processing flow within the artificial intelligence system, and/or augmenting the set of inputs to the artificial intelligence system.) the method comprising:
using at least one sufficiency criteria, determining that the training dataset is insufficient for training the neural network NN; (Cella Paragraph [0525] [0525] In some embodiments, a user of the information technology system may input a modeling goal into the machine learning model 3000. The machine learning model 3000 may learn to analyze training data to output suggestions to the user of the information technology system regarding which types of sensor data are most relevant to achieving the modeling goal, such as one or more types of sensors positioned in, on, or near a value chain entity or a plurality of value chain entities that is relevant to the achievement of the modeling goal is and/or are not sufficient for achieving the modeling goal, and how a different configuration of the types of sensors, such as by adding, removing, or repositioning sensors, may better facilitate achievement of the modeling goal by the machine learning model 3000 and the digital twin system 1700. (Analyzing training data to determine if a defined outcome is feasible. If not, determining that the training data is insufficient))
selecting a generative convolutional neural network GCNN for which the existing training dataset is sufficient to get trained; (Cella Paragraph [0616] Examples of learning algorithms/models include (e.g., deep neural networks, convolution neural networks, and many others as described throughout this disclosure), statistical models (e.g., regression-based models and many others), decision trees and other decision models, random/hidden forests, Hidden Markov models, Bayesian models, and the like. Paragraph [0772] At 5202, a plurality of streams of machine related data from multiple data sources are received at the machine twin 1770. This includes machine specifications like mechanical properties, data from maintenance records, operating data collected from the sensors, historical data including failure data from multiple machines running at different times and under different operating conditions and so on. At 5205, the raw data is cleaned by removing any missing or noisy data, which may occur due to any technical problems in the machine at the time of collection of data. At 5208, one or more models are selected for training by machine twin 1770. The selection of model is based on the kind of data available at the machine twin 1770 and the desired outcome of the model. For example, there may be cases where failure data from machines is not available, or only a limited number of failure datasets exist because of regular maintenance being performed. Classification or regression models may not work well for such cases and clustering models may be most suitable. As another example, if the desired outcome of the model is determining current condition of the machine and detecting any faults, then fault detection models may be selected, whereas if the desired outcome is predicting future failures then remaining useful life prediction model may be selected. At 5210, the one or more models are trained using training dataset and tested for performance using testing dataset. At 5212, the trained model is used for detecting faults and predicting future failure of the machine on production data. Paragraph [0774] Machine twin 1770 then coordinates with artificial intelligence system to select one or more of models based on the kind and quality of available data and the desired answers or outcomes. For example, physical models 5320 may be selected if the intended use of machine twin 1770 is to simulate what-if scenarios and predict how the machine will behave under such scenarios.( Since the selection of the model to be trained is determined based on the type of data available, this is considered to be equivalent to selecting a model for which there is sufficient data to be trained.))
Cella fails to disclose: gaining access to a training dataset for training a neural network NN; training the generative convolutional neural network GCNN using the training dataset or a subset of the training dataset; generating, by the GCNN trained on the existing training dataset, an additional item; and adding the additional item to the training dataset.
In the same field of endeavor, Bradley teaches: gaining access to a training dataset for training a neural network NN; (Col. 9 Lines 29 - 32 As shown, in step 402, training engine 122 trains an encoder neural network and a decoder neural network based on a training dataset that includes multiple sequences of geometries. (In order to use a dataset for training, intuitively access must already be granted.))
training the generative convolutional neural network GCNN using the training dataset or a subset of the training dataset; (Col. 13 Lines 56 - 65 After parameters of discriminator 510 have been updated over one or more epochs, training engine 132 can train generator 500 based on a generator loss 518 that is calculated based on the frequency with which discriminator 510 incorrectly classifies training images 508 from generator 500 as coming from discriminator training data 516. After parameters of generator 500 have been updated over one or more epochs, training engine 132 can resume training discriminator 510 using additional training images 508 produced by generator 500. (Training generator))
generating, by the GCNN trained on the existing training dataset, an additional item; and adding the additional item to the training dataset. (Col. 13 Lines 26 - 38 During training of generator 500, training engine 132 uses generator blocks 602 and/or other components of generator 500 to generate training textures 502(1)-502(M) for various portions of a given synthetic geometry in generator training data 514, where M is an integer greater than one. Next, training engine 132 uses training texture maps 528 for the synthetic geometry to generate screen-space samples 504(1)-504(M) of training textures 502(1)-502(M). Training engine 132 also uses one or more training segmentation masks 530 for the synthetic geometry to generate composited features 506 that include samples 504 that are arranged and/or layered within a single screen-space “image.” Training engine 132 then uses one or more convolutional layers 606 in generator 500 to convert composited features 506 into a training image (e.g., training images 508) in RGB space. (Generating textures for training using a convolutional layer))
It would have been obvious to a person having ordinary skill in the art before the effective filing date to have incorporated the concept of gaining access to a training data set, training a neural network based on that dataset or a subset of that dataset, and generating new items to a training dataset for training a neural network as taught by Bradley into the reference of Cella. Doing so would improve the method of Cella as collecting a large amount of relevant real data can be time and resource intensive, and implementation of this would expedite the process. Certain types of data may not be collected in a quick or efficient manner, and the creation of synthetic data allows for the supplementation of such datasets. (Bradley Col. 1 Lines 31-57)
Regarding Claim 2, the combination of Cella and Bradley teaches the invention as claimed in Claim 1 including: wherein the training dataset comprises items, item labels, andother [sic] meta-information corresponding to each of the items. (Bradley Col. 13 Lines 6 - 23 Returning to the discussion of FIG. 5, training engine 132 trains generator 500 using generator training data 514 that includes training texture maps 528 and training segmentation masks 530 associated with a number of synthetic geometries 526. Synthetic geometries 526 include 3D models of synthetic objects that are similar to objects for which images 540 are to be generated. For example, synthetic geometries 526 could include full-head 3D models of synthetic faces. Training engine 132 and/or another component could generate each synthetic face by randomizing the identity, expression, hairstyle, and/or pose of a parametric face model, such as the face model of FIG. 6B. The component could then generate one or more training texture maps 528 and/or one or more training segmentation masks 530 for each synthetic face by posing and rendering the corresponding face model, as described above with respect to FIGS. 6B-6C. (Training data which has a corresponding element, in this case synthetic faces corresponding to a face model))
Regarding Claim 3, the combination of Cella and Bradley teaches the invention as claimed in Claim 1 including: identifying if the training dataset is sufficient for training the neural network NN based on a sufficiency criterion. (Cella Paragraph [0525] In some embodiments, a user of the information technology system may input a modeling goal into the machine learning model 3000. The machine learning model 3000 may learn to analyze training data to output suggestions to the user of the information technology system regarding which types of sensor data are most relevant to achieving the modeling goal, such as one or more types of sensors positioned in, on, or near a value chain entity or a plurality of value chain entities that is relevant to the achievement of the modeling goal is and/or are not sufficient for achieving the modeling goal, and how a different configuration of the types of sensors, such as by adding, removing, or repositioning sensors, may better facilitate achievement of the modeling goal by the machine learning model 3000 and the digital twin system 1700.(Analyzing training data to determine if a defined outcome is feasible. If not, determining that the training data is insufficient))
Regarding Claim 4, the combination of Cella and Bradley teaches the invention as claimed in Claim 1 including: wherein the GCNN has the same degree of freedom as NN. (Cella Paragraph [0531] In some embodiments, the machine learning model 3000 may be and/or include an artificial neural network, e.g., a connectionist system configured to “learn” to perform tasks by considering examples and without being explicitly programmed with task-specific rules. One or more of the units and/or nodes and connections therebetween may have one or more numerical “weights” assigned. The assigned weights may be configured to facilitate learning, i.e., training, of the machine learning model 3000. The weights assigned weights may increase and/or decrease one or more signals between one or more units and/or nodes, and in some embodiments may have one or more thresholds associated with one or more of the weights. The one or more thresholds may be configured such that a signal is only sent between one or more units and/or nodes, if a signal and/or aggregate signal crosses the threshold. (Degrees of freedom (DoF) in neural networks generally refer to the number of independent parameters (weights and biases) that can be adjusted to minimize error, representing the model's flexibility to fit data.))
Regarding Claim 5, the combination of Cella and Bradley teaches the invention as claimed in Claim 1 including: wherein the degrees of freedom of the GCNN is decreased from the degree of freedom of NN to require fewer items in a GCNN training set to make the existing training set sufficient for the GCNN training. (Cella Paragraph [0531] In some embodiments, the machine learning model 3000 may be and/or include an artificial neural network, e.g., a connectionist system configured to “learn” to perform tasks by considering examples and without being explicitly programmed with task-specific rules. One or more of the units and/or nodes and connections therebetween may have one or more numerical “weights” assigned. The assigned weights may be configured to facilitate learning, i.e., training, of the machine learning model 3000. The weights assigned weights may increase and/or decrease one or more signals between one or more units and/or nodes, and in some embodiments may have one or more thresholds associated with one or more of the weights. The one or more thresholds may be configured such that a signal is only sent between one or more units and/or nodes, if a signal and/or aggregate signal crosses the threshold. Paragraph [0747] An advantage of generating data through simulations and then training machine learning algorithms on this data is the control this approach provides on the features in the data as well as volume and frequency of data. (Degrees of freedom (DoF) in neural networks generally refer to the number of independent parameters (weights and biases) that can be adjusted to minimize error, representing the model's flexibility to fit data.))
Regarding Claim 6, the combination of Cella and Bradley teaches the invention as claimed in Claim 5 including: wherein different parameters comprising degrees of freedom are omitted in the GCNN to generate consecutive new items. (Cella Paragraph [0746] At 5106, one or more stress scenarios may be simulated by changing one or more parameters beyond the normal operating values. The simulating of stress scenarios overcome the limitation of any analysis based only on historical data and helps analyze the network performance across a range of hypothetical yet plausible stress conditions. The simulation involves varying (shocking) one or more parameters while keeping the other parameters as fixed to analyze the impact of such variations on value chain network. In embodiments, a single parameter may be varied while keeping remaining parameters as fixed. In other embodiments, multiple parameters may be varied simultaneously. At 5108, the outcomes of stress scenario simulations are determined, and the performance of value chain network and its different subsystems is estimated across various scenarios. At 5110, the data, parameters and outcomes are fed into a machine learning process in the artificial intelligence system 1160 for further analysis. Paragraph [0747] An advantage of generating data through simulations and then training machine learning algorithms on this data is the control this approach provides on the features in the data as well as volume and frequency of data. (Degrees of freedom (DoF) in neural networks generally refer to the number of independent parameters (weights and biases) that can be adjusted to minimize error, representing the model's flexibility to fit data. As these are independent parameters that can be adjusted, they qualify. An example of a variation in these parameters would be being made equal or equivalent to zero, therefore equivalent to an omission))
Regarding Claim 7, the combination of Cella and Bradley teaches the invention as claimed in Claim 6 including: wherein the omitted degrees of freedom in the GCNN are determined based on statistical parameters related to different parameters of items within the training set. (Cella Paragraph [0745] At 5102, all historical and current data related to the value chain network are received. The data may include information related to various operating parameters of the value chain network over a particular historical time period, say last 12 months. Paragraph [0746] At 5106, one or more stress scenarios may be simulated by changing one or more parameters beyond the normal operating values. The simulating of stress scenarios overcome the limitation of any analysis based only on historical data and helps analyze the network performance across a range of hypothetical yet plausible stress conditions. The simulation involves varying (shocking) one or more parameters while keeping the other parameters as fixed to analyze the impact of such variations on value chain network. Paragraph [0748] In embodiments, the platform may include a system for learning on a training set of outcomes, parameters, and data collected from data sources relating to a set of value chain network activities to train an artificial intelligence/machine learning system to perform stress tests on a physical object using a digital twin that represents a set of value chain entities. (The adjusted parameters relate to differing parameters in the training set, as the adjustments are meant to test the effects on other parameters, therefore forming a relationship))
Regarding Claim 8, the combination of Cella and Bradley teaches the invention as claimed in Claim 1 including: checking the augmented training dataset for sufficiency to train the NN after adding the at least one additional item to the original training set. (Cella Paragraph [2889] In embodiments, the DPLF may be or include the continued process retention of one or more training datasets and/or memories stored in the memory over time. The DPLF thereby allows the ANN to apply existing neural functions and draw upon sets of past events (including ones that are intentionally varied and/or curated for distinct purposes), such as to frame understanding of and behavior within present, recent, and/or new scenarios, including in simulations, during training processes, and in fully operational deployments of the ANN. The DPLF may provide the ANN with a framework by which the ANN may analyze, evaluate, and/or manage data, such as data related to the past, present and future. As such, the DPLF plays a crucial role in training and retraining the ANN via the training system and the retraining system. (Analysis, evaluation, and/or management of training data for retraining purposes.))
Regarding Claim 9, the combination of Cella and Bradley teaches the invention as claimed in Claim 8 including: when checking the augmented training dataset for sufficiency to train the NN, and determining that the augmented training dataset is insufficient to train the NN, executing the method of claim 1. (Cella Paragraph [2890] In embodiments, the DPLF is configured to perform a dual-process operation to manage existing training processes and is also configured to manage and/or perform new training processes, i.e., retraining processes. In embodiments, each instance of the ANN is trained via the training system and configured to be retrained via the retraining system. The ANN encodes training and/or retraining datasets, stores the datasets, and retrieves the datasets during both training via the training system and retraining via the retraining system. The DPANN system 20000 may recognize whether a dataset (the term dataset in this context optionally including various subsets, supersets, combinations, permutations, elements, metadata, augmentations, or the like, relative to a base dataset used for training or retraining), storage activity, processing operation and/or output, has characteristics that natively favor the training system versus the retraining system based on its respective inputs, processing (e.g., based on its structure, type, models, operations, execution environment, resource utilization, or the like) and/or outcomes (including outcome types, performance requirements (including contextual or dynamic requirements), and the like. For example, the DPANN system 20000 may determine that poor performance of the training system on a classification task may indicate a novel problem for which the training of the ANN was not adequate (e.g., in type of data set, nature of input models and/or feedback, quantity of training data, quality of tagging or labeling, quality of supervision, or the like), for which the processing operations of the ANN are not well-suited (e.g., where they are prone to known vulnerabilities due to the type of neural network used, the type of models used, etc.), and that may be solved by engaging the retraining system to retrain the model to teach the model to learn to solve the new classification problem (e.g., by feeding it many more labeled instances of correctly classified items (Recognizing sufficiency and retraining of model))
Regarding Claim 10, the combination of Cella and Bradley teaches the invention as claimed in Claim 8 including: when checking the augmented training dataset for sufficiency to train the NN, and determining that the augmented training dataset is sufficient to train the NN, training the NN. (Cella Paragraph [2890] In embodiments, the DPLF is configured to perform a dual-process operation to manage existing training processes and is also configured to manage and/or perform new training processes, i.e., retraining processes. In embodiments, each instance of the ANN is trained via the training system and configured to be retrained via the retraining system. The ANN encodes training and/or retraining datasets, stores the datasets, and retrieves the datasets during both training via the training system and retraining via the retraining system. The DPANN system 20000 may recognize whether a dataset (the term dataset in this context optionally including various subsets, supersets, combinations, permutations, elements, metadata, augmentations, or the like, relative to a base dataset used for training or retraining), storage activity, processing operation and/or output, has characteristics that natively favor the training system versus the retraining system based on its respective inputs, processing (e.g., based on its structure, type, models, operations, execution environment, resource utilization, or the like) and/or outcomes (including outcome types, performance requirements (including contextual or dynamic requirements), and the like. For example, the DPANN system 20000 may determine that poor performance of the training system on a classification task may indicate a novel problem for which the training of the ANN was not adequate (e.g., in type of data set, nature of input models and/or feedback, quantity of training data, quality of tagging or labeling, quality of supervision, or the like), for which the processing operations of the ANN are not well-suited (e.g., where they are prone to known vulnerabilities due to the type of neural network used, the type of models used, etc.), and that may be solved by engaging the retraining system to retrain the model to teach the model to learn to solve the new classification problem (e.g., by feeding it many more labeled instances of correctly classified items). (Recognizing sufficiency and retraining/training of model)).
Regarding Claims 11 – 19, they are system claims that correspond to the method claims 1 - 10 above. Therefore, they are rejected for the same reason as method claims 1 - 10 above.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Pathak et al. (US 12481873 B1) is a method and system for generative design based on deep learning.
MATSUO (US 20250165799 A1) discusses training data generation.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JUSTIN A CARDOSO whose telephone number is (571)272-8512. The examiner can normally be reached M-F 7:30 - 5:00, alternate Friday's off.
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, Jennifer Welch can be reached at (571) 272-7212. 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.
/JUSTIN CARDOSO/
Patent Examiner, Art Unit 2143
/JENNIFER N WELCH/Supervisory Patent Examiner, Art Unit 2143
Prosecution Insights