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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on September 03, 2026 has been entered. Claims 1-5, 7-15, and 17-20 are pending in the application.
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 1-5, 7-15, and 17-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed towards abstract ideas without significantly more.
Claim 1 interpretation: Under the broadest reasonable interpretation (BRI), the terms of the claim are presumed to have their plain meaning consistent with the specification as it would be interpreted by one of ordinary skill in the art. See MPEP 2111. Based on the specification, the recitation “generate reaction data as a function of the collected physiological responses of the user” (see specification ¶[0033]-[0034]) is being interpreted as judgements/observations (the labeling of events) and/or mathematical calculations/evaluations (the classifier). The recitation “generate a plurality of process data of a process, wherein the plurality of process data further comprises a plurality of process input data and a plurality of correlated process output data” (see specification ¶[0003]-[0004] the correlation of the input data to output data; and see specification ¶[0042], the (pre)processing and/or feature extraction) is being interpreted as judgements/observations and/or mathematical calculations/evaluations. The recitation “performing feature extraction on the physiological responses of the user” (see specification ¶[0042]) is being interpreted as judgements/observations and/or mathematical calculations/evaluations. The recitation “generating the cognitive data as a function of a cognition classifier” (see specification ¶[0032] and ¶[0080]) is being interpreted as judgements/observations (the labeling of events) and/or mathematical calculations/evaluations (the classifier). The recitation “determining the reaction data as a function of a user input and the cognitive data” (see specification ¶[0034]) is being interpreted as mathematical calculations/evaluations. The recitation “generate a first process model using the plurality of process data, wherein generating the first process model further comprises training the process model using the plurality of process input data and the plurality of correlated process output data, wherein the first process model comprises the cognitive classifier” (see specification ¶[0025]) is being interpreted as mathematical calculations/evaluations. The recitation “train, using the plurality of metamodel training examples, a metamodel, wherein the metamodel is configured to receive process model measurements and output model efficiency scores, wherein the metamodel comprises a process metamodel configured to receive the cognition classifier and the historical data as inputs” (see specification ¶[0026]-[0027]) is being interpreted as mathematical calculations/evaluations. The recitation “generate a measurement of the first process model” (see specification ¶[0028]) is being interpreted as mathematical calculations/evaluations. The recitation “identify, using the process metamodel, an incorrect output of the cognition classifier by comparing a first association of a mental state to an event generated by the cognition classifier with the association of the mental state to the event included in the historical data” (see specification ¶[0036]-[0037]) is being interpreted as mathematical calculations/evaluations. The recitation “retrain the cognition classifier with training data including the association of the mental state to the event included in the historical data” (see specification ¶[0037] and ¶[0067]-[0068]). The recitation “identify a collateral process as a function of the metamodel, wherein the collateral process indicates an additional output identified by the metamodel as a function of the plurality of metamodel training examples correlated to historical data” (see specification ¶[0038]) is being interpreted as mathematical calculations/evaluations. The recitations are computer-implemented, as indicated in the specification (see ¶[0009]-[0010], ¶[0023], ¶[0042], and ¶[0070]), and in the claim, lines 4-6.
Claim 11 interpretation: Under the broadest reasonable interpretation (BRI), the terms of the claim are presumed to have their plain meaning consistent with the specification as it would be interpreted by one of ordinary skill in the art. See MPEP 2111. Based on the specification, the recitation “generating a plurality of process data of a process, wherein the plurality of process data further comprises a plurality of process input data and a plurality of correlated process output data” (see specification ¶[0003]-[0004] the correlation of the input data to output data; and see specification ¶[0042], the (pre)processing and/or feature extraction) is being interpreted as judgements/observations and/or mathematical calculations/evaluations. The recitation “performing feature extraction on the physiological responses of the user” (see specification ¶[0042]) is being interpreted as judgements/observations and/or mathematical calculations/evaluations. The recitation “generating the cognitive data as a function of a cognition classifier” (see specification ¶[0032] and ¶[0080]) is being interpreted as judgements/observations (the labeling of events) and/or mathematical calculations/evaluations (the classifier). The recitation “determining the reaction data as a function of a user input and the cognitive data” (see specification ¶[0034]) is being interpreted as mathematical calculations/evaluations. The recitation “generating a first process model using the plurality of process data, wherein generating the first process model further comprises training the process model using the plurality of process input data and the plurality of correlated process output data, wherein the first process model comprises the cognition classifier” (see specification ¶[0025]) is being interpreted as mathematical calculations/evaluations. The recitation “training a metamodel using the plurality of metamodel training examples, wherein the metamodel is configured to receive process model measurements and output model efficiency scores, wherein the metamodel comprises a process metamodel configured to receive the cognition classifier and the historical data as inputs” (see specification ¶[0026]-[0027]) is being interpreted as mathematical calculations/evaluations. The recitation “generating a measurement of the first process model” (see specification ¶[0028]) is being interpreted as mathematical calculations/evaluations. The recitation “identifying, using the process metamodel, an incorrect output of the cognition classifier by comparing a first association of a mental state to an event generated by the cognition classifier with the association of the mental state to the event included in the historical data” (see specification ¶[0036]-[0037]) is being interpreted as mathematical calculations/evaluations. The recitation “retraining the cognition classifier with training data including the association of the mental state to the event included in the historical data” (see specification ¶[0037] and ¶[0067]-[0068]). The recitation “identifying a collateral process as a function of the metamodel, wherein the collateral process indicates an additional output identified by the metamodel as a function of the plurality of metamodel training examples correlated to historical data” (see specification ¶[0038]) is being interpreted as mathematical calculations/evaluations. The recitations are computer-implemented, as indicated in the specification (see ¶[0009]-[0010], ¶[0023], ¶[0042], and ¶[0070]), and in the claim, line 4.
Step 1: This part of eligibility analysis evaluates whether the claim falls within any statutory category. MPEP 2106.03. Claim 1 is directed towards an apparatus for identifying collateral processes, which is directed towards a machine and/or a manufacture (a statutory category of invention). Claim 11 recites a method for identifying collateral processes, which is directed towards a process (a statutory category of invention). Step 1: YES.
Step 2A Prong One: This part of the eligibility analysis evaluates whether the claim recites a judicial exception. As explained in MPEP 2106.04(a)(2)(I). The courts consider mathematical calculations, when the claim is given its BRI in light of the specification, as falling within the “mathematical concept” grouping of abstract ideas. A claim does not have to recite “calculating” in order to be considered a mathematical calculation. For example, a step of “determining” a variable or number using a mathematical method, or “performing” a mathematical operation, may also be considered a mathematical calculation when the BRI of the claim in light of the specification encompasses a mathematical calculation. As discussed in the claim interpretation section, the limitations include, under the BRI, various mathematical calculations/evaluations. Accordingly, the limitations as seen in claims 1 and 11 recite judicial exceptions (abstract ideas that fall within the mathematical calculations grouping of mathematical concepts).
Alternatively or additionally, these steps describe the concept of using implicit mathematical formulas (i.e., calculations to determine a likelihood score) to derive a conclusion based on input of data, which corresponds to concepts identified as abstract ideas by the courts (Diamond v. Diehr. 450 U.S. 175, 209 U.S.P.Q. 1 (1981), Parker v. Flook. 437 U.S. 584, 19 U.S.P.Q. 193 (1978), and In re Grams. 888 F.2d 835, 12 U.S.P.Q.2d 1824 (Fed. Cir. 1989)). The concept of the recited limitations identified as mathematical concepts above is not meaningfully different than those mathematical concepts found by the courts to be abstract ideas.
Furthermore, as explained in MPEP 2106.04(a)(2)(III). The courts consider a mental process (thinking) that “can be performed in the human mind, or by a human using a pen and paper” to be an abstract idea. CyberSource Corp. v. Retail Decisions, Inc., 654 F.3d 1366, 1372, 99 USPQ2d 1690, 1695 (Fed. Cir. 2011). The “mental processes” abstract idea grouping is defined as concepts performed in the human mind, and examples of mental processes include observations, evaluations, judgements, and opinions. As discussed in the claim interpretation section, the limitations include, under the BRI, judgements/observations and/or mathematical calculations/evaluations. Accordingly, the limitations as seen in claims 1 and 11 recite judicial exceptions (abstract ideas that fall within the mental process grouping). No limitations are provided that would force the complexity of any of the identified evaluation steps to be non-performable by pen-and-paper practice.
In particular, claim 1 recites the following elements, which are part of the abstract idea (i.e., the algorithm):
generate reaction data as a function of the collected physiological responses of the user;
generate a plurality of process data of a process as a function of outputs from the plurality of sensors, wherein the plurality of process data further comprises a plurality of process input data and a plurality of correlated process output data, wherein the plurality of process input data comprises cognitive data correlated to the reaction data, wherein the cognitive data comprises complexity data, wherein the complexity data comprises cognitive data related to overcoming a problem, wherein the complexity data further comprises a cognitive effect of at least a step to solve the problem, wherein cognitive data further comprises severity data, wherein the severity data comprises cognitive data related to at least an effect of the problem on a brain of the user, wherein generating the plurality of process data further comprises:
performing feature extraction on the physiological responses of the user collected using the plurality of sensors,
generating the cognitive data as a function of a cognition classifier; and
determining the reaction data as a function of a user input and the cognitive data;
generate a first process model using the plurality of process data, wherein generating the first process model further comprises training the process model using the plurality of process input data and the plurality of correlated process output data, wherein the first process model comprises the cognition classifier;
receive a plurality of metamodel training examples, wherein each metamodel training example of the plurality of metamodel training examples includes a plurality of model input examples and model output examples corresponding to an exemplary model, and wherein each model training example within the plurality of metamodel training examples further includes at least an efficiency score per the exemplary model relating to the plurality of model input examples and model output examples, wherein the exemplary model comprises a prior iteration of the cognition classifier, wherein the plurality of model input examples comprises prior inputs of the cognition classifier, wherein the model output examples comprise prior outputs of the cognition classifier, and wherein the plurality of metamodel training examples comprises historical data including an association of a mental state to an event;
train, using the plurality of metamodel training examples, a metamodel, wherein the metamodel is configured to receive process model measurements and output model efficiency scores, wherein the metamodel comprises a process metamodel configured to receive the cognition classifier and the historical data as inputs;
generate a measurement of the first process model;
output a model efficiency score of the first process model using the metamodel and the measurement of the first process model;
identify, using the process metamodel, an incorrect output of the cognition classifier by comparing a first association of a mental state to an event generated by the cognition classifier with the association of the mental state to the event included in the historical data;
retrain the cognition classifier with training data including the association of the mental state to the event included in the historical data;
identify a collateral process as a function of the metamodel, wherein the collateral process indicates an additional output identified by the metamodel as a function of the plurality of metamodel training examples correlated to historical data;
transmit the collateral process.
Furthermore, claim 11 recites the following elements, which are part of the abstract idea (i.e., the algorithm):
a method for identifying collateral processes, the method comprising:
generating a plurality of process data of a process, wherein the plurality of process data further comprises a plurality of process input data and a plurality of correlated process output data, wherein the plurality of process input data comprises cognitive data correlated to reaction data, wherein the cognitive data comprises complexity data, wherein the complexity data comprises cognitive data related to overcoming a problem, wherein generating the plurality of process data further comprises:
performing feature extraction on the physiological responses of the user collected using the plurality of sensors,
generating the cognitive data as a function of a cognition classifier; and
determining the reaction data as a function of a user input and the cognitive data;
generating a first process model using the plurality of process data, wherein generating the first process model further comprises training the process model using the plurality of process input data and the plurality of correlated process output data, wherein the first process model comprises the cognition classifier;
receiving a plurality of metamodel training examples, wherein each metamodel training example of the plurality of metamodel training examples includes a plurality of model input examples and model output examples corresponding to an exemplary model, and wherein each metamodel training example within the plurality of metamodel training examples further includes at least an efficiency score per the exemplary model relating to the plurality of model input examples and model output examples, wherein the exemplary model comprises a prior iteration of the cognition classifier, wherein the plurality of model input examples comprises prior inputs of the cognition classifier, wherein the model output examples comprise prior outputs of the cognition classifier, and wherein the plurality of metamodel training examples comprises historical data including an association of a mental state to an event;
training a metamodel using the plurality of metamodel training examples, wherein the metamodel is configured to receive process model measurements and output model efficiency scores, wherein the metamodel comprises a process metamodel configured to receive the cognition classifier and the historical data as inputs;
generating a measurement of the first process model;
outputting a model efficiency score of the first process model using the metamodel and the measurement of the first process model;
identifying, using the process metamodel, an incorrect output of the cognition classifier by comparing a first association of a mental state to an event generated by the cognition classifier with the association of the mental state to the event included in the historical data;
retraining the cognition classifier with training data including the association of the mental state to the event included in the historical data;
identifying a collateral process as a function of the metamodel, wherein the collateral process indicates an additional output identified by the metamodel as a function of the plurality of metamodel training examples correlated to historical data;
transmitting the collateral process.
Step 2A Prong One: YES.
Step 2A Prong Two: This part of the eligibility analysis evaluates whether the claim as a whole integrates the judicial exceptions into a practical application of the exception. This evaluation is performed by (a) identifying whether there are any additional elements recited in the claim beyond the judicial exceptions, and (b) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exceptions into a practical application. Claims 1 and 11 recite the additional elements of a processor, memory, and a user device (see specification ¶[0031] the user device is a generic computer, ¶[0046] the user interface is a GUI on the user device). The apparatus/method are merely instructions to implement an abstract idea on a generic computer or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.04(d) and MPEP 2106.05(f).
The claims further recite an additional element of a plurality of sensors. The plurality of sensors does not qualify as integration into a practical application because this limitation is merely adding insignificant extra-solution activity to the judicial exception, i.e., mere data gathering at a higher level of generality - see MPEP 2106.04(d) and MPEP 2106.05(g) using generic components (i.e., the sensors are generic).
Step 2A Prong Two: NO.
Step 2B: This part of the eligibility analysis evaluates whether the claim as a whole, amounts to significantly more than the recited exception, i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. MPEP 2106.05. As explained with Step 2A Prong Two, the claims recite additional elements which are directed towards the usage of a generic computer, and are at best the equivalent of merely adding the words “apply it” to the judicial exceptions. Mere instructions to apply an exception cannot provide an inventive concept. These elements/steps can be seen as well-understood, routine, and conventional individually and in combination. Claims 1 and 11 recite the additional elements of a processor, memory, and a user device (see specification ¶[0031] the user device is a generic computer, ¶[0046] the user interface is a GUI on the user device). Receiving further user input merely requires receiving data, which is just a function of a generic computer, and/or insignificant data gathering. Thus, the apparatus/method do not qualify as significantly more because these limitations are simply appending well-understood, routine and conventional activities previously known in the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer functions that are well-understood, routine and conventional activities previously known in the industry (see Electric Power Group, 830 F.3d 1350 (Fed. Cir. 2016); Alice Corp. v. CLS Bank Int’l, 110 USPQ2d 1976 (2014)) and/or a claim to an abstract idea requiring no more than being stored on a computer readable medium which is a well-understood, routine and conventional activity previously known in the industry (see Electric Power Group, 830 F.3d 1350 (Fed. Cir. 2016); Alice Corp. v. CLS Bank Int’l, 110 USPQ2d 1976 (2014); SAP Am. v. InvestPic, 890 F.3d 1016 (Fed. Circ. 2018)).
The claims further recite an additional element of a plurality of sensors. The plurality of sensors does not qualify as significantly more because (1) this is simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer functions that are well-understood, routine and conventional activities previously known to the industry and/or (2) this limitation is merely adding insignificant extra-solution activity to the judicial exception, i.e., mere data gathering at a higher level of generality - see MPEP 2106.04(d) and MPEP 2106.05(g) using generic components (i.e., the sensors are generic). In this case, as the sensors are claimed generically, they may be considered generic computer components.
Alternatively and/or additionally, if the sensors were to include an eye-tracking sensor, Tomasi et al. (US Patent Application Publication 2019/0046029 – cited in prior action) teaches imaging the eyes of a subject (see abstract) utilizing conventional mobile device hardware, including the built-in camera (see ¶[0010]), so that different metrics may be obtained, including the movement of the eyes (see ¶[0039] and ¶[0111]-[0113]; Figs. 10A-11). Therefore, the plurality of sensors cannot be seen as significantly more.
The claims also recite a cognition classifier which is merely insignificant extrasolution activity applied to the judicial exception, e.g., mere data manipulation using a well-known element (i.e., the classifier) claimed generically. Machine learning, when claimed generically, is well-known, routine, and conventional in the art. For example, see Hu et al. (“Intelligent Sensor Networks The Integration of Sensor Networks, Signal Processing and Machine Learning”, CRC Press, 2012/10/23) teaches machine learning may utilize data from wireless sensor networks in supervised and unsupervised fashion (see pg. 3-7), and may be utilized for predictions (see pg. 153-180). See Huang et al. (“Kernal Based Algorithms for Mining Huge Data Sets Supervised, Semi-supervised, and Unsupervised Learning”, Springer, 2006) teaches about supervised learning, including for predicting labels of unseen data (see pg. 1-3). See Mitchell (“The Discipline of Machine Learning”, Machine Learning Department, Carnegie Mellon, July 2006) teaches about the state of machine learning, and including the usage in biological settings (see pg. 1-7). Therefore, the cognition classifier cannot be seen as significantly more.
Looking at the limitations as an ordered combination (that is, as a whole) adds nothing that is not already present when looking at the elements taking individually. There is no indication that the combination of elements improves the functioning of a computer, for example, or improves any other technology. There is no indication that the combination of elements permits automation of specific tasks that previously could not be automated. There is no indication that the combination of elements includes a particular solution to a computer-based problem or a particular way to achieve a desired computer-based outcome. Rather, the collective functions of the claimed invention merely provide conventional computer implementation, i.e., the computer is simply a tool to perform the process. Step 2B: NO.
Claims 1 and 11 are not eligible.
Claims 2-5 and 7-10; and 12-15 and 17-20 depend from claims 1 and 11, respectively, and merely further define the abstract ideas of claims 1 and 11 with no further element that integrates the abstract ideas into a practical application or that qualifies as being significantly more. Looking at the limitations of each claim as an ordered combination in conjunction with the claims from which they depend (that is, as a whole) adds nothing that is not already present when looking at the elements taken individually. There is no indication that the combination of elements improves the functioning of a computer, for example, or improves any other technology. There is no indication that the combination of elements permits automation of specific tasks that previously could not be automated. There is no indication that the combination of elements includes a particular solution to a computer-based problem or a particular way to achieve a desired computer-based outcome.
Claims 7 and 17 further contemplate a wearable device; however, the specification of the present application does not contemplate what the wearable device could be (see specification ¶[0081]). Therefore, the examiner is interpreting the wearable device as a generic smartphone which could be worn on a user. Therefore, the apparatus/method are merely instructions to implement an abstract idea on a generic computer (i.e., the generic smartphone) or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.04(d) and MPEP 2106.05(f). Further that, the claims recite additional elements which are directed towards the usage of a generic computer (i.e., the generic smartphone), and are at best the equivalent of merely adding the words “apply it” to the judicial exceptions. Mere instructions to apply an exception cannot provide an inventive concept. These elements/steps can be seen as well-understood, routine, and conventional individually and in combination. Thus, the apparatus/method do not qualify as significantly more because these limitations are simply appending well-understood, routine and conventional activities previously known in the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer functions that are well-understood, routine and conventional activities previously known in the industry (see Electric Power Group, 830 F.3d 1350 (Fed. Cir. 2016); Alice Corp. v. CLS Bank Int’l, 110 USPQ2d 1976 (2014)) and/or a claim to an abstract idea requiring no more than being stored on a computer readable medium which is a well-understood, routine and conventional activity previously known in the industry (see Electric Power Group, 830 F.3d 1350 (Fed. Cir. 2016); Alice Corp. v. CLS Bank Int’l, 110 USPQ2d 1976 (2014); SAP Am. v. InvestPic, 890 F.3d 1016 (Fed. Circ. 2018)).
Response to Arguments
Applicant’s arguments, 35 U.S.C. § 112(a)
Applicant’s arguments, see pg. 1, filed September 03, 2026, with respect to the rejections of claims 1-5, 7-15, and 17-20 under 35 U.S.C. § 112(a) have been fully considered and are persuasive. Therefore, the rejections have been withdrawn.
Applicant’s arguments, 35 U.S.C. § 101
Applicant’s arguments, see pg. 1-12, filed September 03, 2026, with respect to the rejections of claims 1-5, 7-15, and 17-20 under 35 U.S.C. § 101 have been fully considered and are NOT persuasive. Applicant first argues that, with respect to Step 2A, prong one, the position that the pending claims do not recite a mental process or mathematical calculation is maintained. The examiner respectfully disagrees for the reasons as set forth in the prior actions. Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that the claims, as a whole, integrate the judicial exceptions into a practical application by reciting “the particular process-metamodel and retraining steps that provide the disclosed improvement to operation of the cognition classifier”. The examiner respectfully disagrees.
In order to be an improvement, any improvement must be disclosed “first the specification should be evaluated to determine if the disclosure provides sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement” and that “if the specification sets forth an improvement in technology, the claim must be evaluated to ensure that the claim itself reflects the disclosed improvement. That is, the claim includes the components or steps of the invention that provide the improvement described in the specification.”
The specification details that “As used herein, “process metamodel,” is a model configured to improve accuracy, efficiency and interpretability of outputs of machine learning models described in this disclosure” and “At each iteration of the comparison of first process model 124 to second process model 160 using metamodel 152, the functionality of apparatus 100 including processor 104 may be improved at least due to the improvements in the accuracy of the output generated by metamodel 152” (see ¶[0036] and ¶[0030]); and that “any process of training, retraining, deployment, and/or instantiation of any machine-learning model and/or algorithm may be performed and/or repeated after an initial deployment and/or instantiation to correct, refine, and/or improve the machine-learning model and/or algorithm” (see specification ¶[0067]-[0068]). Such an improvement is therefore directed towards merely improving the accuracy of the output of the model. As such, Applicant’s improvement does not lie in the functioning of the computer, rather, the improvement is directed towards the abstract ideas themselves (i.e., the improvement in the machine learning output of the cognition classifier, the identification of a collateral process).
An improved mental process is still a mental process even if such a mental process results in more accurate results.1,2 Also, having the claims focus on determining the collateral process is not itself limiting the claims to improving the technology because cases that involve practical, technological improvements extend beyond simply improving the accuracy of a prediction.3 See, e.g., McRO, Inc. v. Bandai Namco Games America Inc., 837 F.3d 1299, 1315 (Fed. Cir. 2016) (“The claimed process uses a combined order of specific rules that renders information into a specific format that is then used and applied to create desired results: a sequence of synchronized, animated characters.”); Finjan, Inc. v. Blue Coat Sys., Inc., 879 F.3d 1299, 1304 (Fed. Cir. 2018) (finding patent eligible a claim drawn to a behavior-based virus scan that protects against viruses that have been “cosmetically modified to avoid detection by code-matching virus scans”); Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1330, 1333 (Fed. Cir. 2016) (discussing patent eligible claims directed to “an innovative logical model for a computer database” that included a self-referential table allowing for greater flexibility in configuring databases, faster searching, and more effective storage); CardioNet, LLC v. InfoBionic, Inc., 955 F.3d 1358, 1368 (Fed. Cir. 2020) (explaining that the claims at issue focus on a specific means for improving cardiac monitoring technology; they are not “directed to a result or effect that itself is the abstract idea and merely invoke generic processes and machinery” (quoting McRO, 837 F.3d at 1314)). Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that the claim now “recites a corrective retraining operation that changes how the cognition classifier is trained after the process metamodel identifies a historically inconsistent learned correlation. That is the relevant portion of similarity to Desjardins… both claims recite a concrete, disclosed mechanism that alters the operation and subsequent training of a machine-learning model based on historically derived model information”. The examiner respectfully disagrees.
The improvements in Desjardins provided an “improvement to machine learning technology by explaining how the machine learning model is trained to learn new tasks while protecting knowledge about previous tasks to overcome the problem of “catastrophic forgetting,” and that the claims reflected the improvement identified in the specification… allowing the system to reduce use of storage capacity; and the enablement of reduced complexity in the system” and that “[s]uch improvements were tantamount to how the machine learning model itself would function in operation and therefore not subsumed in the identified mathematical calculation” (see December 05, 2025 USPTO Memo, revised second and last paragraphs of MPEP § 2106.04(d)(1)).
The improvement in Desjardins provides multiple key distinctions as compared to the present application; primarily, the improvements in Desjardins improved the functioning of a computer (i.e., allowing the system to reduce use of storage capacity; and the enablement of reduced complexity in the system) and the improvement in Desjardins were not subsumed in the identified mathematical calculation (i.e., the identified abstract ideas). Conversely, the improvement of the claims of the present application does not improve any computer function, and lies in the machine learning output (i.e., improving accuracy from the retraining) and the identification of a collateral process. As described above, any such improvements of the present application are directed towards abstract ideas themselves. Such that, in the present application as compared to Desjardins, such improvements are subsumed in the identified abstract ideas. Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that, with regard to Example 39, “Although Example 39 was found not to recite a judicial exception at Prong One, it remains instructive because it illustrates that a specific iterative correction-and-retraining architecture is not merely a result-oriented instruction to use machine learning”. The examiner respectfully disagrees. The analysis of Example 39 does not discuss a specific iterative correction-and-retraining architecture to not being a result-oriented instruction to use machine-learning. It is not apparent how Example 39 relates to the claim at issue, as the claim at issue recites judicial exceptions, and the claim of Example 39 does not. Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that, as opposed to Example 47, claim 2, the amended claim includes additional details (that explain the analysis of detected anomalies, from Example 47, claim 2), that “claim 1 does not merely invoke a cognition classifier or metamodel as a technological environment in which abstract evaluation occurs. It recites the particular corrective training process by which the classifier is modified in response to historically identified error”. The examiner respectfully disagrees. While the claim of the present application is not commensurate in scope with claim 2 of Example 47, the claim of the present application similar uses a generic computer to implement recited exceptions. There is nothing particular of the computer utilized in the present application, similar to claim 2 of Example 47. Furthermore, the claims at issue confines the claim to a particular technological environment. The additional detail of retraining the model is particular machine learning environment, and do not add any inventive concept to the claim. Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that the claim as amended, when considered as a whole, “recites components and steps that integrate any alleged exception into a practical application… The newly added process-metamodel and retraining limitations are not generic processor, display, sensor, or user-interface functions. They define the particular interaction between historical data, the process metamodel, the cognition classifier, and retraining data that corrects an identified learned correlation in the classifier itself”. The examiner respectfully disagrees. All of the machine learning related steps (i.e., the interaction between historical data, the process metamodel, the cognition classifier, and retraining data) are implemented via a generic computer, and are generic computer processes from a generic processor. There is nothing about this claim that would prohibit implementation on any generic computer, so it is not clear how the claim could not be implemented via a generic computer. Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that the claim as amended “no longer rests on an alleged general assertion that a metamodel produces a more accurate score or identifies a more useful collateral process. Rather, the claim recites a particular classifier-correction architecture in which historically inconsistent classifier behavior is identified and incorporated into subsequent training of that same classifier”. Further that amended claim 1 “now reflects the disclosed improvement through the particular historical-error identification and classifier-retraining steps”. The examiner respectfully disagrees. As explained above, the specification details that “As used herein, “process metamodel,” is a model configured to improve accuracy, efficiency and interpretability of outputs of machine learning models described in this disclosure” and “At each iteration of the comparison of first process model 124 to second process model 160 using metamodel 152, the functionality of apparatus 100 including processor 104 may be improved at least due to the improvements in the accuracy of the output generated by metamodel 152” (see ¶[0036] and ¶[0030]); and that “any process of training, retraining, deployment, and/or instantiation of any machine-learning model and/or algorithm may be performed and/or repeated after an initial deployment and/or instantiation to correct, refine, and/or improve the machine-learning model and/or algorithm” (see specification ¶[0067]-[0068]). Such an improvement is therefore directed towards merely improving the accuracy of the output of the model. Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that, with regard to In re Board of Trustees of Leland Stanford Junior University, “Applicant does not assert eligibility merely because the cognition classifier yields a more accurate classification, a different reaction score, or a more informative collateral-process output… The process metamodel identifies an incorrect learned correlation in the cognition classifier in view of a historical association, and the processor retrains the cognition classifier using training data including that historical association. The claimed advance is therefore not limited to storing, displaying, or reporting a model result. Rather, claim 1 specifies how historically identified classifier error changes subsequent training of the classifier itself”. The examiner disagrees. The process of identifying an incorrect output of the cognition classifier, and then retraining the cognition classifier, does not improve any technical process, but merely improves the accuracy of the cognition classifier output (i.e., corrected for incorrect outputs), which is similar to the claim at issue in In re Board of Trustees of Leland Stanford Junior University, as even if the enhancement to the abstract idea (i.e., retraining the cognition classifier) yields different or better results, this does not render patent eligible subject matter. Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that contrast to Synopsys, Inc. v. Mentor Graphics Corp, “claim 1 expressly recites an apparatus including sensors, a processor, and memory, and further recites a particular process-metamodel correction mechanism. The newly added limitations do not merely aid a person in mentally identifying a relationship. They require generation and use of historical metamodel training examples, identification of an incorrect cognition-classifier input, output, or correlation using the process metamodel, and retraining of the cognition classifier using historical training data. That is a claimed way of correcting the trained operation of the classifier, not the claimed idea of identifying a correlation.” The examiner respectfully disagrees. While Applicant’s claim recite additional elements including sensors, a processor, and memory, these are not enough to integrate the judicial exceptions into a practical application. As explained above, the additional elements of a processor, memory, and a user device (see specification ¶[0031] the user device is a generic computer, ¶[0046] the user interface is a GUI on the user device). The apparatus/method are merely instructions to implement an abstract idea on a generic computer or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.04(d) and MPEP 2106.05(f).
The claims further recite an additional element of a plurality of sensors. The plurality of sensors does not qualify as integration into a practical application because this limitation is merely adding insignificant extra-solution activity to the judicial exception, i.e., mere data gathering at a higher level of generality - see MPEP 2106.04(d) and MPEP 2106.05(g) using generic components (i.e., the sensors are generic). Furthermore, the “particular process-metamodel correction mechanism” is merely instructions executing on a processor, directed towards the abstract ideas. Therefore, Applicant’s arguments are not persuasive.
Next, Applicant argues that “distinguishable from the claims addressed in Electric Power Group and SAP Am. v. InvestPic,. Claim 1 does not merely collect information, analyze that information, and present a resulting collateral process. Rather, claim 1 requires the process metamodel to identify an incorrect cognition-classifier association in view of historical data and retrain the cognition classifier using training data containing the historical association. These limitations recite a particular mechanism for correcting training classifier behavior and specify how the claimed system uses historically identified error in subsequent classifier training. Accordingly, claim 1 is not directed to a generalized workflow of collecting, analyzing, and displaying information.” The examiner respectfully disagrees. The claims require no more than being stored on a computer readable medium which is a well-understood, routine and conventional activity previously known in the industry (see Electric Power Group, 830 F.3d 1350 (Fed. Cir. 2016); Alice Corp. v. CLS Bank Int’l, 110 USPQ2d 1976 (2014); SAP Am. v. InvestPic, 890 F.3d 1016 (Fed. Circ. 2018)), as the cognition-classifier elements are all well-known computer implementations, operable on a generic computer. The claims also recite the sensors, but as described above, the plurality of sensors does not qualify as integration into a practical application because this limitation is merely adding insignificant extra-solution activity to the judicial exception, i.e., mere data gathering at a higher level of generality - see MPEP 2106.04(d) and MPEP 2106.05(g) using generic components (i.e., the sensors are generic). Therefore, Applicant’s arguments are not persuasive.
Accordingly, the claim when considered as a whole does not integrate the judicial exceptions into a practical application.
Next, Applicant argues that, with regard to Step 2B, “amended claim 1 recites significantly more than any alleged judicial exception because it requires a particular, ordered classifier-correction arrangement, not a generic instruction to apply machine learning to cognitive data”. The examiner respectfully disagrees. As explained above, the claims recite additional elements which are directed towards the usage of a generic computer, and are at best the equivalent of merely adding the words “apply it” to the judicial exceptions. Mere instructions to apply an exception cannot provide an inventive concept. These elements/steps can be seen as well-understood, routine, and conventional individually and in combination. Claims 1 and 11 recite the additional elements of a processor, memory, and a user device (see specification ¶[0031] the user device is a generic computer, ¶[0046] the user interface is a GUI on the user device). Receiving further user input merely requires receiving data, which is just a function of a generic computer, and/or insignificant data gathering. Thus, the apparatus/method do not qualify as significantly more because these limitations are simply appending well-understood, routine and conventional activities previously known in the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer functions that are well-understood, routine and conventional activities previously known in the industry (see Electric Power Group, 830 F.3d 1350 (Fed. Cir. 2016); Alice Corp. v. CLS Bank Int’l, 110 USPQ2d 1976 (2014)) and/or a claim to an abstract idea requiring no more than being stored on a computer readable medium which is a well-understood, routine and conventional activity previously known in the industry (see Electric Power Group, 830 F.3d 1350 (Fed. Cir. 2016); Alice Corp. v. CLS Bank Int’l, 110 USPQ2d 1976 (2014); SAP Am. v. InvestPic, 890 F.3d 1016 (Fed. Circ. 2018)).
The claims further recite an additional element of a plurality of sensors. The plurality of sensors does not qualify as significantly more because (1) this is simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer functions that are well-understood, routine and conventional activities previously known to the industry and/or (2) this limitation is merely adding insignificant extra-solution activity to the judicial exception, i.e., mere data gathering at a higher level of generality - see MPEP 2106.04(d) and MPEP 2106.05(g) using generic components (i.e., the sensors are generic). In this case, as the sensors are claimed generically, they may be considered generic computer components.
Alternatively and/or additionally, if the sensors were to include an eye-tracking sensor, Tomasi et al. (US Patent Application Publication 2019/0046029 – cited in prior action) teaches imaging the eyes of a subject (see abstract) utilizing conventional mobile device hardware, including the built-in camera (see ¶[0010]), so that different metrics may be obtained, including the movement of the eyes (see ¶[0039] and ¶[0111]-[0113]; Figs. 10A-11). Therefore, the plurality of sensors cannot be seen as significantly more.
The claims also recite a cognition classifier which is merely insignificant extrasolution activity applied to the judicial exception, e.g., mere data manipulation using a well-known element (i.e., the classifier) claimed generically. Machine learning, when claimed generically, is well-known, routine, and conventional in the art. For example, see Hu et al. (“Intelligent Sensor Networks The Integration of Sensor Networks, Signal Processing and Machine Learning”, CRC Press, 2012/10/23) teaches machine learning may utilize data from wireless sensor networks in supervised and unsupervised fashion (see pg. 3-7), and may be utilized for predictions (see pg. 153-180). See Huang et al. (“Kernal Based Algorithms for Mining Huge Data Sets Supervised, Semi-supervised, and Unsupervised Learning”, Springer, 2006) teaches about supervised learning, including for predicting labels of unseen data (see pg. 1-3). See Mitchell (“The Discipline of Machine Learning”, Machine Learning Department, Carnegie Mellon, July 2006) teaches about the state of machine learning, and including the usage in biological settings (see pg. 1-7). Therefore, the cognition classifier cannot be seen as significantly more.
Looking at the limitations as an ordered combination (that is, as a whole) adds nothing that is not already present when looking at the elements taking individually. There is no indication that the combination of elements improves the functioning of a computer, for example, or improves any other technology. There is no indication that the combination of elements permits automation of specific tasks that previously could not be automated. There is no indication that the combination of elements includes a particular solution to a computer-based problem or a particular way to achieve a desired computer-based outcome. Rather, the collective functions of the claimed invention merely provide conventional computer implementation, i.e., the computer is simply a tool to perform the process. Therefore, Applicant’s arguments are not persuasive.
As such, Applicant’s arguments with respect to Step 2B are not persuasive. For commensurate reasons, Applicant’s arguments are not persuasive with respect to claim 11. The dependent claims remain rejected based on their dependence from a rejected base claim. Therefore, Applicant’s arguments are not persuasive and the rejections to the claims under 35 U.S.C. § 101 are maintained.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JONATHAN D. MORONESO whose telephone number is (571)272-8055. The examiner can normally be reached M-F: 8:30AM - 6:00 PM, MST.
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 M. ROBERTSON can be reached at (571)272-5001. 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.
/J.D.M./Examiner, Art Unit 3791
/JENNIFER ROBERTSON/Supervisory Patent Examiner, Art Unit 3791
1 “[T]he improvement in computational accuracy alleged here does not qualify as an improvement to a technological
process; rather, it is merely an enhancement to the abstract mathematical calculation of haplotype phase itself...The
different use of a mathematical calculation, even one that yields different or better results, does not render patent
eligible subject matter.” In re Board of Trustees of Leland Stanford Junior University, 991 F.3d 1245 (Fed. Cir.
2021).
2 “[A] claim for a new abstract idea is still an abstract idea.” Synopsys, Inc. v. Mentor Graphics Corp, 839 F.3d 1138
(Fec. Cir. 2016).
3 See In re Board of Trustees of Leland Stanford Junior University, 991 F.3d 1245 (Fed. Cir. 2021).
