DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
This action is responsive to the claims filed 7/10/2026. Claims 1-2 are amended. Claims 1-7 are pending.
Specification
The disclosure is objected to because of the following informalities:
(a) At page 2, line 5 of the Substitute Specification filed 9/20/2023, the specification recites "the fitness test being defined based on a set of input/output pairs each indicating the required output of the function for an input". The test introduced in the immediately preceding clause is "a verification test", and claim 1 as amended recites "the verification test being defined based on a set of input/output pairs each indicating a required output of the function for an input". The term "the fitness test" has no antecedent basis in the specification and is inconsistent with the terminology used throughout the remainder of the disclosure and in the claims. It appears that "the fitness test" should read "the verification test".
(b) At page 3, lines 4 through 6 of the Substitute Specification filed 9/20/2023, the paragraph beginning "According to a third aspect of the present disclosure" is identical to the second aspect paragraph at page 3, lines 1 through 3 but for the ordinal, and therefore sets out no aspect distinct from the second aspect. The same duplication appears in the international application as filed, so it was not introduced by the Substitute Specification and no third aspect was lost. Claim 7 is directed to a non-transitory computer-readable storage, and the Summary states no aspect corresponding to that subject matter. The brief summary of the invention should be consistent with the subject matter of the claims. See 37 CFR 1.73 and MPEP 608.01(d). Correction of the third aspect paragraph, consistent with claim 7 and without the introduction of new matter, is required.
(c) At page 3, lines 1 and 4, the specification recites "there is a provided a computer system". The phrase is grammatically incorrect and should read "there is provided a computer system".
(d) At page 3, line 11, the specification recites "Figure 1 is a block diagram a computer system suitable for the operation of embodiments of the present disclosure". The word "of" appears to have been omitted. Compare page 3, line 19, which recites "Figure 1 is a block diagram of a computer system suitable for the operation of embodiments of the present disclosure".
Appropriate correction is required.
Claim Objections
Claim 7 is objected to because of the following informalities:
Claim 7 recites "A non-transitory computer-readable storage comprising computer program code to, when loaded into a computer system and executed thereon, cause the computer system to perform the method as claimed in claim 1". Claim 7 is presented with the status identifier "(Previously Presented)", but the claim of record, as amended by the preliminary amendment filed 9/20/2023 and as quoted in the Office action mailed 5/6/2026, reads "A non-transitory computer-readable storage medium comprising computer program code to, when loaded into a computer system and executed thereon, cause the computer system to perform the method as claimed in claim 1". The word "medium" has been omitted from the claim as presented, without the status identifier "Currently Amended" and without the markings required by 37 CFR 1.121(c). The omission appears to be inadvertent, and claim 7 should read "A non-transitory computer-readable storage medium comprising computer program code to, when loaded into a computer system and executed thereon, cause the computer system to perform the method as claimed in claim 1".
Appropriate correction is required.
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 1-7 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The analysis of the claims will follow the 2019 Revised Patent Subject Matter Eligibility Guidance, 84 Fed. Reg. 50 (“2019 PEG”).
Claim 1
Step 1: This claim recites “A computer implemented method…comprising”; therefore, it is directed to the statutory category of a process.
Step 2A Prong 1: This claim recites, inter alia:
model a function comprising: defining a verification test for verifying that the AI algorithm models the function, the verification test being defined based on a set of input/output pairs each indicating a required output of the function for an input: These limitations recite a mathematical relationship by further organizing information and manipulating information, e.g. a computer implemented method of deploying an artificial intelligence (AI) algorithm to model a function comprising: defining a verification test for verifying that the AI algorithm models the function, through mathematical correlations. e.g. the verification test being defined based on a set of input/output pairs each indicating a required output of the function for an input.
defining a machine learning component having a machine learning algorithm and a configuration, the machine learning algorithm being trained based on training data to model the function: Furthers the mathematical relationship by further organizing and manipulating information, e.g. defining a machine learning component having a machine learning algorithm and a configuration, through mathematical correlations, e.g. the machine learning algorithm being trained based on training data to model the function.
iteratively adapting the machine learning component over a plurality of generations, wherein each generation of the component is adapted by modifying the configuration of the component, and wherein the adaptation for a generation is selected from a set of candidate adaptations based on a determination of a fitness of the component so adapted, the fitness being determined by the verification test: These limitations further the mathematical relationship by applying it repeatedly and manipulating information on its result. For each candidate adaptation, the component as so adapted is related by the verification test to the outputs the input/output pairs require, a fitness is determined from that relation, and the adaptation for the generation is selected on what that relation yields; the fitness so determined need be no number, the claim is satisfied where the verification test returns only a bare verdict that the component as adapted does or does not produce the required outputs, and such a verdict is itself the result of relating the produced outputs to the required outputs, which is the correlation the claim recites (MPEP 2106.04(a)(2)(I); operative specification page 5, lines 16 to 17, and page 5, lines 17 to 20, which describe the fitness determination as binary in some embodiments and as a measure in others); the selected adaptation then supplies the configuration to which the same relation is applied at the next generation. Organizing and manipulating information through mathematical correlations in this way is a mathematical concept (MPEP 2106.04(a)(2)(I); Digitech, 758 F.3d at 1350). The claim requires no particular configuration, no particular number of candidate adaptations, no particular number of generations, no particular fitness function, and no ranking or comparison across all candidates.
wherein the iterative adaptation ceases in response to a stopping condition such that, on cessation, a latest generation of the machine learning component is selected to constitute the AI algorithm modelling the function: These limitations recite no act outside the concept already recited. It states the condition on which the recited iteration of the correlation comes to an end, and designates the last result of that same correlation as the answer. Where the stopping condition is a measure of fitness reaching a predetermined threshold, one of the two forms claim 4 names, the condition is itself the comparison of the value the verification test yields against a number, which is the manipulating of information through mathematical correlations (MPEP 2106.04(a)(2)(I)); where it is a count of iterations, it does no more than state how many times the recited relation is applied. The claim recites no act of deploying, installing, loading, compiling or executing the resulting AI algorithm; the last event the claim recites is that a latest generation "is selected to constitute" the AI algorithm. Applicant's disclosure treats these as steps of the same method rather than as anything added to it: "At 306 the method iteratively adapts a configuration of the machine learning component over multiple generations until a stopping condition is reached. At 308 a latest generation of the machine learning component is selected as the AI algorithm 214" (operative specification page 6, lines 23-26).
Thus, this claim recites a judicial exception.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The additional elements of this claim are as follows:
A computer implemented method of deploying an artificial intelligence (AI) algorithm to: These addition elements are recited at the highest level of generality and amounts to mere instructions to apply the judicial exception using a generic computer as a tool (MPEP 2106.05(f)). Applicant's own specification establishes the generic character of that computer: "A central processor unit (CPU) 102 is communicatively connected to a storage 104 and an input/output (I/O) interface 106 via a data bus 108. The storage 104 can be any read/write storage device such as a random-access memory (RAM) or a non-volatile storage device... Examples of I/O devices connectable to I/O interface 106 include a keyboard, a mouse, a display (such as a monitor) and a network connection" (operative specification page 3, lines 20-26), and the method is described as implementable "using a software-controlled programmable processing device, such as a microprocessor, digital signal processor or other processing device, data processing apparatus or system" (operative specification page 6, lines 28-29). No particular machine is recited (MPEP 2106.05(b)), nothing is transformed (MPEP 2106.05(c)), and the element imposes no other meaningful limit on practicing the exception (MPEP 2106.05(e)).
Thus, the way in which the additional elements use or interact with the judicial exception do not integrate the judicial exception into a practical application when this claim is considered as a whole.
Step 2B: The additional elements from Step 2A Prong 2 include invoking a computer or other machinery to apply the underlying judicial exception. Thus, the additional elements, viewed individually or in combination, do not provide an inventive concept or otherwise amount to significantly more than the abstract idea itself. See MPEP 2106.05.
Claim 2
Step 1: a process, as in claim 1.
Step 2A Prong 1: This claim recites, inter alia:
wherein multiple machine learning components are defined, each having a machine learning algorithm and a configuration trained based on training data to model the function: These limitations recite a mathematical relationship of organizing information and manipulating information, e.g. wherein multiple machine learning components are defined, each having a machine learning algorithm and a configuration, through mathematical correlations. e.g. trained based on training data to model the function. "trained based on training data" is again participial and recites no act of training.
and wherein the iterative adapting is performed for each of the multiple machine learning components, the method further comprising: responsive to the stopping condition, a measure of fitness of each of the latest generation of each machine learning component is compared to select a fittest machine learning component to constitute the AI algorithm modelling the function: Furthers the mathematical relationship by further organizing and manipulating information, e.g. and wherein the iterative adapting is performed for each of the multiple machine learning components, the method further comprising: responsive to the stopping condition, a measure of fitness of each of the latest generation of each machine learning component, through mathematical correlations, e.g. is compared to select a fittest machine learning component to constitute the AI algorithm modelling the function.
Thus, this claim recites a judicial exception.
Step 2A Prong 2 & Step 2B: There are no additional elements recited so this claim does not provide a practical application and is not considered to be significantly more. As such, the claim is patent ineligible.
Claim 3
Step 1: a process, as in claim 1.
Step 2A Prong 1: This claim recites the same judicial exception as in claim 1.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The additional elements of this claim are as follows:
wherein the configuration of the machine learning component includes one or more of: a layer depth; a neuron function; an adjustment factor; or an adjustment function: These additional elements are recited at a high level of generality and represent generally linking the underlying abstract idea to a field of use or technological environment. See MPEP 2106.05(h).
Thus, the way in which the additional elements use or interact with the judicial exception do not integrate the judicial exception into a practical application when this claim is considered as a whole.
Step 2B: The additional elements from Step 2A Prong 2 include generally linking the underlying abstract idea to a field of use or technological environment. Thus, the additional elements, viewed individually or in combination, do not provide an inventive concept or otherwise amount to significantly more than the abstract idea itself. See MPEP 2106.05.
Claim 4
Step 1: a process, as in claim 1.
Step 2A Prong 1: This claim recites the same judicial exception as in claim 1.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The additional elements of this claim are as follows:
wherein the stopping condition includes one or more of: a predetermined number of iterations; or a predetermined threshold measure of fitness determined by the verification test: These additional elements are recited at a high level of generality and represent generally linking the underlying abstract idea to a field of use or technological environment. See MPEP 2106.05(h).
Thus, the way in which the additional elements use or interact with the judicial exception do not integrate the judicial exception into a practical application when this claim is considered as a whole.
Step 2B: The additional elements from Step 2A Prong 2 include generally linking the underlying abstract idea to a field of use or technological environment. Thus, the additional elements, viewed individually or in combination, do not provide an inventive concept or otherwise amount to significantly more than the abstract idea itself. See MPEP 2106.05.
Claim 5
Step 1: a process, as in claim 1.
Step 2A Prong 1: This claim recites the same judicial exception as in claim 1.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The additional elements of this claim are as follows:
wherein at least a subset of the set of candidate adaptations is generated randomly: These additional elements are recited at a high level of generality and represent generally linking the underlying abstract idea to a field of use or technological environment. See MPEP 2106.05(h).
Thus, the way in which the additional elements use or interact with the judicial exception do not integrate the judicial exception into a practical application when this claim is considered as a whole.
Step 2B: The additional elements from Step 2A Prong 2 include generally linking the underlying abstract idea to a field of use or technological environment. Thus, the additional elements, viewed individually or in combination, do not provide an inventive concept or otherwise amount to significantly more than the abstract idea itself. See MPEP 2106.05.
Claim 6
Step 1: This claim is directed to “A computer system”; therefore, this claim is directed to the statutory category of machines.
Step 2A Prong 1: This claim recites the same judicial exception as in claim 1.
Step 2A Prong 2: The judicial exception recited in this claim is not integrated into a practical application. The only substantive difference between claim 6 and claim 1 is that claim 6 is directed to “A computer system comprising a processor and memory storing computer program code for performing the method”. However, mere recitation that a judicial exception is to be performed using generic computer components in their ordinary capacity cannot meaningfully integrate the judicial exception into a practical application. See MPEP 2106.05(f). With that exception, the analysis at this step for claim 6 mirrors that of claim 1.
Step 2B: The additional elements from Step 2A Prong 2 do not contain significantly more than the judicial exception for this claim. The only substantive difference between claim 6 and claim 1 is that claim 6 is directed to “A computer system comprising a processor and memory storing computer program code for performing the method”. However, mere recitation that a judicial exception is to be performed using generic computer components in their ordinary capacity cannot amount to significantly more than the judicial exception. See MPEP 2106.05(f). With that exception, the analysis at this step for claim 6 mirrors that of claim 1.
Claim 7
Step 1: This claim is directed to “A non-transitory computer-readable storage medium”; therefore, this claim is directed to an article of manufacture.
Step 2A Prong 1: This claim recites the same judicial exception as in claim 1.
Step 2A Prong 2: The judicial exception recited in this claim is not integrated into a practical application. The only substantive difference between claim 7 and claim 1 is that claim 6 is directed to “A non-transitory computer-readable storage medium comprising computer program code to, when loaded into a computer system and executed thereon, cause the computer system to perform the method”. However, mere recitation that a judicial exception is to be performed using generic computer components in their ordinary capacity cannot meaningfully integrate the judicial exception into a practical application. See MPEP 2106.05(f). With that exception, the analysis at this step for claim 7 mirrors that of claim 1.
Step 2B: The additional elements from Step 2A Prong 2 do not contain significantly more than the judicial exception for this claim. The only substantive difference between claim 7 and claim 1 is that claim 6 is directed to “A non-transitory computer-readable storage medium comprising computer program code to, when loaded into a computer system and executed thereon, cause the computer system to perform the method”. However, mere recitation that a judicial exception is to be performed using generic computer components in their ordinary capacity cannot amount to significantly more than the judicial exception. See MPEP 2106.05(f). With that exception, the analysis at this step for claim 7 mirrors that of claim 1.
Claim Rejections - 35 USC 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-7 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by McDonnell et al. (hereinafter McDonnell), US 2020/0175378 A1.
Regarding independent claim 1, McDonnell discloses a computer implemented method of deploying an artificial intelligence (AI) algorithm to model a function comprising (McDonnell: [0028], "a processor executes the automated model generation process 120"; [0011], the system "may be implemented using (e.g., executed by) one or more computing devices, such as laptop computers, desktop computers, mobile devices, servers", which is the recited computer implementation; [0017], the resulting "one or more models 124 are executable by the processor that executes the automated model generation process 120 (or by another processor or by another device) to perform an operation, such as classification, clustering, anomaly detection, or some other type of operation based on input data"; [0029], "The one or more models 124 are trained to perform a task based on input data", so the AI algorithm that issues from the method models the recited function):
defining a verification test for verifying that the AI algorithm models the function (McDonnell: [0053], "The genetic algorithm 410 includes or is otherwise associated with a fitness function 440, a stagnation criterion 450, a crossover operation 460, and a mutation operation 470. The fitness function 440 is an objective function that can be used to compare the models of the input set 420", which is the recited verification test; [0066], "The output data of each model may be evaluated using the fitness function 440 to determine how well the model modeled the input data set 402", which is the recited verifying that the AI algorithm models the function), the verification test being defined based on a set of input/output pairs each indicating a required output of the function for an input (McDonnell: [0053], "In some examples, the fitness function 440 is based on a frequency and/or magnitude of errors produced by testing a model on the input data set 402. As a simple example, assume the input data set 402 includes ten rows, that the input data set 402 includes two columns denoted A and B, and that the models illustrated in FIG. 4 represent neural networks that output a predicted value of B given an input value of A. In this example, testing a model may include inputting each of the ten values of A from the input data set 402, comparing the predicted values of B to the corresponding actual values of B from the input data set 402, and determining if and/or by how much the two predicted and actual values of B differ"; each of the ten (A, B) rows indicates the required output B of the function for the input A, and the fitness function defined on those rows is the recited verification test);
defining a machine learning component having a machine learning algorithm and a configuration (McDonnell: [0050], "Each member of the input set 420 and the output set 430 is a model (e.g., a data structure) that represents a neural network"; [0062], "The genetic algorithm 410 may automatically generate an initial set of models based on the input data set 402 and configuration input", which is the recited defining; [0065], "The genetic algorithm 410 may automatically assign an activation function, an aggregation function, a bias, connection weights, etc. to each model of the input set 420 for the initial epoch"; [0051], "each model may specify a neural network by at least a neural network topology, a series of activation functions, and connection weights" and "The models may also be specified to include other parameters, including but not limited to bias values/functions and aggregation functions", which is the recited configuration), the machine learning algorithm being trained based on training data to model the function (McDonnell: [0088], "sending the trainable model to an optimization trainer, and adding a trained model received from the optimization trainer as an input to a second epoch of the genetic algorithm that is subsequent to the first epoch", and "at least one model (e.g., the trainable model 422) may be provided to the backpropagation trainer 480 for training into the trained model 482. The fitness function may be evaluated based on the input data set, and the optimization trainer may be configured to use a portion of the input data set to train the trainable model", the portion so used being the recited training data; [0029], "the one or more models 124 may be trained based on the input data set 102 to perform a classification task"; [0080], "the automated model generation process 120 is executed to generate and train the one or more models 124"); and
iteratively adapting the machine learning component over a plurality of generations (McDonnell: [0050], "During each iteration of the search process (also called an "epoch" or "generation" of the genetic algorithm 410), an input set 420 (or population) is "evolved" to generate an output set 430 (or population)" and "the output set 430 of one epoch may be the input set 420 of the next (non-initial) epoch"; [0076], "The genetic algorithm 410 may continue in the manner described above through multiple epochs", which is the recited plurality of generations), wherein each generation of the machine learning component is adapted by modifying the configuration of the machine learning component (McDonnell: [0071], "The rest of the output set 430 may be filled out by random reproduction using the crossover operation 460 and/or the mutation operation 470"; [0072], the crossover operation produces "a neural network that differs from each of its "parent" neural networks in terms of topology, activation function, aggregation function, bias value/function, link weight, or any combination thereof"; [0075], "the mutation operation 470 may cause the topology of a particular model of the input set to be modified by addition or omission of one or more input nodes, by addition or omission of one or more connections, by addition or omission of one or more hidden nodes, or a combination thereof. As another example, the mutation operation 470 may cause one or more activation functions, aggregation functions, bias values/functions, and/or or connection weights to be modified"; each item so modified is an element of the configuration recited at [0051]; and the operators act on the individual model, which is the recited machine learning component, [0073], the crossover operation 460 "generates a model of the output set 430 by combining aspects of a first model of the input set 420 with aspects of one or more other models of the input set 420", and [0075], the mutation operation 470 "generates or contributes to a model of the output set 430 by mutating any aspect of a model of the input set 420"; so the model that constitutes the component at each successive generation is one whose configuration has been modified), and wherein the adaptation for a generation is selected from a set of candidate adaptations based on a determination of a fitness of the machine learning component so adapted (McDonnell: [0071], "The rest of the output set 430 may be filled out by random reproduction using the crossover operation 460 and/or the mutation operation 470", and [0070], "In the illustrated example, the output set 430 includes the same number of models as the input set 420", so the models so produced in an epoch are a plurality of differently adapted models, a set of candidate adaptations rather than one; [0071], "After the output set 430 is generated, the output set 430 may be provided as the input set 420 for the next epoch of the genetic algorithm 410", and [0063], "the output set 430 generated during the initial epoch may become the input set 420 of the next epoch of the genetic algorithm 410", so the models next tested are the components as adapted; [0066], "Each model of the input set 420 may be tested based on the input data set 402 to determine model fitness", which is the recited determination of a fitness of the machine learning component so adapted; [0069], "The fittest models of each "elite species" may be identified. The fittest models overall may also be identified"; [0070], "The output set 430 may include each of the "overall elite" models and each of the "elite member" models. Propagating the "overall elite" and "elite member" models to the next epoch may preserve the "genetic traits" resulted in caused such models being assigned high fitness values", so the adaptation carried into the succeeding generation is the one selected out of that set on the determined fitness, and because the input set 420 of an epoch is the output set 430 of the epoch before it ([0071], [0063]), the model so tested, so ranked and so carried forward is a model the crossover operation 460 or the mutation operation 470 produced by modifying the configuration in the preceding epoch, so that one and the same model answers this limitation and the preceding one; McDonnell states the same fitness governed mechanic in summary at [0088], "based on a fitness function, selecting, by the processor, a subset of models from the plurality of models (the plurality of models based on a genetic algorithm and corresponding to a first epoch of the genetic algorithm), performing, by the processor, at least one genetic operation of the genetic algorithm with respect to at least one model of the subset to generate a trainable model"), the fitness being determined by the verification test (McDonnell: [0066], "The output data of each model may be evaluated using the fitness function 440 to determine how well the model modeled the input data set 402"; the fitness so determined comes from the same fitness function 440 of [0053] that is the verification test mapped above; [0067], "a species fitness may be determined for each of the species. The species fitness of a species may be a function of the fitness of one or more of the individual models in the species"),
wherein the iterative adaption ceases in response to a stopping condition such that, on cessation, a latest generation of the machine learning component is selected to constitute the AI algorithm modelling the function (McDonnell: [0076], "The genetic algorithm 410 may continue in the manner described above through multiple epochs until a specified termination criterion, such as a time limit, a number of epochs, or a threshold fitness value (e.g., of an overall fittest model), is satisfied. When the termination criterion is satisfied, an overall fittest model of the last executed epoch may be selected and output as reflecting the topology of the one or more models 124 of FIG. 1"; the last executed epoch is the recited latest generation and the model selected out of it constitutes the recited AI algorithm; [0016], "the one or more models 124 are the "fittest" models generated during the last epoch of the automated model generation process 120").
Regarding dependent claim 2, McDonnell further discloses the method of claim 1, wherein multiple machine learning components are defined, each having a machine learning algorithm and a configuration trained based on training data to model the function (McDonnell: [0051], "The input set 420 and the output set 430 each includes a plurality of models, where each model includes data representative of a neural network"; [0064], "the input set 420 may include models with multiple distinct topologies. For example, a first model may have a first topology, including a first number of input nodes associated with a first set of data parameters, a first number of hidden layers including a first number and arrangement of hidden nodes, one or more output nodes, and a first set of interconnections between the nodes. In this example, a second model of epoch may have a second topology"; [0065], the algorithm "may automatically assign an activation function, an aggregation function, a bias, connection weights, etc. to each model of the input set 420"; [0088], "at least one model (e.g., the trainable model 422) may be provided to the backpropagation trainer 480 for training into the trained model 482 ... the optimization trainer may be configured to use a portion of the input data set to train the trainable model", the trained model so produced being "add[ed] ... as an input to a second epoch of the genetic algorithm"; [0080], "For example, the automated model generation process 120 is executed to generate and train the one or more models 124"; [0029], "the one or more models 124 may be trained based on the input data set 102 to perform a classification task". McDonnell's express statement of training on a portion of the input data set is of at least one model per epoch, and not in terms of every model of the plurality; the limitation is nevertheless met because what it requires is that each component defined BE one whose machine learning algorithm and configuration are trained on training data to model the function, and every model of the plurality has its configuration derived from, evaluated against and refined on the input data set 402 through the process McDonnell itself calls generating and training the models [0065], [0066], [0080], [0088]), and wherein the iterative adapting is performed for each of the multiple machine learning components (McDonnell: [0050], during each iteration "an input set 420 (or population) is "evolved" to generate an output set 430 (or population)", the whole population being evolved each epoch; [0066], "Each model of the input set 420 may be tested based on the input data set 402 to determine model fitness"; [0070] and [0071], the succeeding output set is composed of the fitness-selected models together with the crossover and mutation products that replace the remainder), the method further comprising: responsive to the stopping condition, a measure of fitness of each of the latest generation of each machine learning component is compared to select a fittest machine learning component to constitute the AI algorithm modelling the function (McDonnell: [0076], "When the termination criterion is satisfied, an overall fittest model of the last executed epoch may be selected and output as reflecting the topology of the one or more models 124 of FIG. 1", which identifies an overall fittest by comparison of the fitness measures across the last executed epoch; [0069], "The fittest models of each "elite species" may be identified. The fittest models overall may also be identified"; [0067], "The genetic algorithm 410 may maintain a data structure that tracks the fitness of each species across multiple epochs").
Regarding dependent claim 3, McDonnell further discloses the method of claim 1, wherein the configuration of the machine learning component includes one or more of: a layer depth; a neuron function; an adjustment factor; or an adjustment function (McDonnell: [0062], "For example, the neural network topology may indicate a number of input nodes, a number of hidden layers, a number of nodes per hidden layer, and a number of output nodes", the number of hidden layers being the recited layer depth; [0052], "The activation function of a node may be a step function, sine function, continuous or piecewise linear function, sigmoid function, hyperbolic tangent function, or another type of mathematical function that represents a threshold at which the node is activated", which is the recited neuron function; [0052], "The connection weight is a numerical quantity that influences if and/or how the output of N1 is modified before being input at N2" and "The bias is a constant value or function that is used by the aggregation function and/or the activation function to make the node more or less likely to be activated", either of which is the recited adjustment factor; [0052], "The aggregation function is a mathematical function that combines (e.g., sum, product, etc.) input signals to the node", which is the recited adjustment function).
Regarding dependent claim 4, McDonnell further discloses the method of claim 1, wherein the stopping condition includes one or more of: a predetermined number of iterations; or a predetermined threshold measure of fitness determined by the verification test (McDonnell: [0076], "a specified termination criterion, such as a time limit, a number of epochs, or a threshold fitness value (e.g., of an overall fittest model), is satisfied", the number of epochs being the recited predetermined number of iterations and the threshold fitness value being the recited predetermined threshold measure of fitness, evaluated with the same fitness function 440 of [0053] that constitutes the verification test; [0061], "a maximum number of epochs the genetic algorithm 410 will be executed, a threshold fitness value that results in termination of the genetic algorithm 410 even if the maximum number of generations has not been reached").
Regarding dependent claim 5, McDonnell further discloses the method of claim 1, wherein at least a subset of the set of candidate adaptations is generated randomly (McDonnell: [0071], "The rest of the output set 430 may be filled out by random reproduction using the crossover operation 460 and/or the mutation operation 470"; [0074], "During the mutation operation 470, a portion of a model is randomly modified" and "the mutation operation 470 may include randomly modifying the value of one or more bits/bytes/portions in a normalized vector"; [0073], "the crossover operation 460 may be a random or pseudo-random operator that generates a model of the output set 430 by combining aspects of a first model of the input set 420 with aspects of one or more other models of the input set 420").
Regarding dependent claim 6, McDonnell further discloses a computer system comprising a processor and memory storing computer program code for performing the method of claim 1 (McDonnell: [0011], "The system 100, or portions thereof, may be implemented using (e.g., executed by) one or more computing devices, such as laptop computers, desktop computers, mobile devices, servers, and Internet of Things devices and other devices utilizing embedded processors and firmware or operating systems"; [0012], "The instructions may be stored at a memory, such as a random-access memory (RAM), a read-only memory (ROM), a computer-readable storage device, an enterprise storage device, any other type of memory, or a combination thereof" and "the operations described with reference to the parameter selector 104 and the automated model generation process 120 are performed by a processor (e.g., a central processing unit (CPU), graphics processing unit (GPU), or other type of processor)"). The substantive limitations of claim 6 are the limitations of claim 1 and are met on the same basis.
Regarding dependent claim 7, McDonnell further discloses a non-transitory computer-readable storage medium comprising computer program code to, when loaded into a computer system and executed thereon, cause the computer system to perform the method as claimed in claim 1 (McDonnell: [0098], "the system may take the form of a computer program product on a computer-readable storage medium or device having computer-readable program code (e.g., instructions) embodied or stored in the storage medium or device. Any suitable computer-readable storage medium or device may be utilized, including hard disks, CD-ROM, optical storage devices, magnetic storage devices, and/or other storage media" and, in the same paragraph, "As used herein, a "computer-readable storage medium" or "computer-readable storage device" is not a signal", which establishes the recited non-transitory character; [0100], "Computer program instructions may be loaded onto a computer or other programmable data processing apparatus to produce a machine, such that the instructions that execute on the computer or other programmable data processing apparatus create means for implementing the functions specified in the flowchart block or blocks"). The substantive limitations of claim 7 are the limitations of claim 1 and are met on the same basis.
Response to Arguments
Applicant's amendment and Remarks filed 7/10/2026 have been fully considered. Every argument Applicant presented is identified and answered below, in the order Applicant presented it.
Regarding the objection to claims 1 and 2.
Applicant's arguments, see page 6 of Remarks filed 7/10/2026, with respect to the objection to claims 1 and 2 set forth in the Office action mailed 5/6/2026 have been fully considered and are persuasive. That objection has been withdrawn.
Regarding the objection to the specification.
Applicant's arguments, see page 6 of Remarks filed 7/10/2026, with respect to the objection to the specification set forth in the Office action mailed 5/6/2026 have been fully considered and are persuasive. The objection to the specification set forth in that action has been withdrawn.
Regarding the rejection of claim 2 under 35 U.S.C. 112(b).
Applicant's arguments, see page 6 of Remarks filed 7/10/2026, with respect to the rejection of claim 2 under 35 U.S.C. 112(b) set forth in the Office action mailed 5/6/2026 have been fully considered and are persuasive. That rejection has been withdrawn.
Regarding the arguments directed to the rejection under 35 U.S.C. 101.
Applicant's arguments filed 7/10/2026 with respect to the rejection of claims 1-7 under 35 U.S.C. 101 have been fully considered but they are not persuasive. The rejection is maintained, on the judicial exception identified in the Office action mailed 5/6/2026.
(1) The argument that claim 1 integrates any abstract idea into a practical application, the specification identifying a technical problem and claim 1 supplying a specific technical solution (Remarks, page 6, lines 23 to 26, and page 7, lines 1 to 7). Applicant argues that "[c]laim 1 is not directed to an abstract idea without significantly more because the method integrates any idea that arguendo may be considered 'abstract' into a practical application, specifically by improving the technology of AI deployment", and that "the method of claim 1 improves the technological process of artificial intelligence deployment". Applicant argues further that paragraph [0004] of the specification "identifies a specific technical problem: selecting a 'good-fit' AI algorithm and its accompanying configuration parameters is challenging and not readily achieved in an automated manner", and that "[t]he method of claim 1 provides a specific technical solution by automating this selection through an iterative, multi-generational adaptation of the machine learning component's configuration, guided by a specific verification test."
Examiner respectfully disagrees. This argument is not persuasive. It is an argument under Step 2A Prong Two that claim 1 integrates the recited judicial exception into a practical application by improving a technology, and it is answered on that footing. The passage relied upon is in the disclosure and is a statement of the problem, not of the claim. Under MPEP 2106.05(a), after the specification has been consulted and the disclosed invention determined to improve technology, "the claim must be evaluated to ensure the claim itself reflects the disclosed improvement in technology"; "the claim must include the components or steps of the invention that provide the improvement described in the specification"; and "[t]he full scope of the claim under the BRI should be considered to determine if the claim reflects an improvement in technology." Applicant identifies a problem described in the disclosure and does not identify any recitation of claim 1 that supplies the asserted solution. Identifying a technical problem in a specification is not reciting a technical solution in a claim.
The word on which the argument turns is "specific", and claim 1 recites nothing specific. Claim 1 recites no particular verification test: at its broadest reasonable scope the recitation reads on any comparison of produced outputs against required outputs. It recites no particular adaptation: any modification to any configuration parameter satisfies it. It recites no particular selection rule: the fitness on which the selection turns may be a bare fit-or-not-fit verdict (Substitute Specification, page 5, lines 16 to 17) rather than a measure (page 5, lines 17 to 20), and ties among equally fit candidates may be broken at random (page 5, lines 22 to 24). That breadth does not take the claim outside the mathematical concepts grouping: a verdict that the component as adapted does or does not produce the required outputs is itself the result of relating the produced outputs to the required outputs, and it is that correlation, not any particular arithmetic, that the claim recites (MPEP 2106.04(a)(2)(I)). And it recites no particular stopping condition: the disclosure's list of stopping conditions is introduced by "for example" and "inter alia" (Substitute Specification, page 6, lines 6 to 9; the amendment to that paragraph filed 7/10/2026 corrects "can includes" to "can include" and leaves the words relied on here unchanged), and claim 4 recites that the stopping condition "includes one or more of" two named conditions. Each of these is recited at its genus. That is the distinction MPEP 2106.05(a) itself draws in discussing McRO, where the particular rules were recited in the claim.
Two matters of record are noted so that the record shows they were not overlooked. First, the paragraph at page 7, lines 1 to 7 of the remarks ends in the middle of a sentence, at "...guided by a specific verification test. T". Whatever further argument was intended is not of record and cannot be answered. Second, Applicant cites throughout to the paragraph numbers of US 2024/0177066, the pre-grant publication of the application as filed. The operative description of record is the Substitute Specification - Clean filed 9/20/2023, as amended 7/10/2026. Both passages Applicant relies upon are present in the operative copy, the [0004] passage at page 1, lines 23 to 25 and the [0003] passage at page 1, lines 14 to 22, and the answers here are given to those loci.
(2) The argument that claim 1 is not the mere automation of a manual process (Remarks, page 7, lines 8 to 16). Applicant argues that the method of claim 1 "is not simple automation of something that could otherwise be done manually, however, at least because the problem being solved does not exist outside of a computing environment"; that "[u]nder MPEP 2106.05(a), an improvement to the functioning of a computer distinguishes a claim from the mere automation of manual processes"; that "[s]electing an architecture for an AI software component is inherently a computer-technology problem"; and that "[b]ecause the method of claim 1 configures and deploys a machine learning algorithm (a software component), it is improving the way the computer system configures its own software".
Examiner respectfully disagrees. This argument is not persuasive, for three reasons. First, MPEP 2106.05(a) is to the opposite effect. That section lists "[m]ere automation of manual processes, such as using a generic computer to process an application for financing a purchase", Credit Acceptance Corp. v. Westlake Services, 859 F.3d 1044, 1055 (Fed. Cir. 2017), and speeding up a loan-application process, LendingTree, LLC v. Zillow, Inc., 656 Fed. App'x 991, 996-97 (Fed. Cir. 2016) (non-precedential), among the "[e]xamples that the courts have indicated may not be sufficient to show an improvement in computer-functionality." The section nowhere provides that a process which could not be performed manually is for that reason an improvement in computer functionality. The proposition Applicant draws from the section is not in it. Second, claim 1 recites no act of selecting an architecture. The only thing any recitation of claim 1 modifies is "the configuration of the machine learning component". The type of machine learning algorithm is fixed by the component as defined and is never adapted; no recitation of claim 1 substitutes one type of machine learning algorithm for another. Claim 3's recitation that the configuration "includes one or more of: a layer depth; a neuron function; an adjustment factor; or an adjustment function" is permissive and, being a recitation of a dependent claim, does not narrow claim 1, as set out in the preliminary matter above. The disclosure's "architectural selections (such as type(s) of algorithm(s) (e.g. regressor or classifier, long-short-term-memory, deep neural network, convolutional neural network etc.), extent of supervision in training data, training technique and the like)" appear in the Background of the disclosure (Substitute Specification, page 1, lines 17 to 19), and the selection of an architecture is named in the title of the application. Neither is a limitation of any claim. Third, claim 1 recites no act of deploying and no act of configuring a computer system, as set out in the preliminary matter above. Nothing recited in claim 1 is installed, loaded into memory, executed, served or run, and no computer system's operation is altered by anything claim 1 recites. The assertion that claim 1 improves "the way the computer system configures its own software" therefore rests on acts the claim does not contain.
This answer does not depend on which recitation of claim 1 is treated as an additional element at Step 2A Prong Two. On any identification of the additional elements, MPEP 2106.05(a) requires that the claim include the components or steps that provide the improvement described in the specification, evaluated at the full scope of the claim under the broadest reasonable interpretation, and Applicant identifies none.
(3) The argument that claim 1 cannot practically be performed in the human mind (Remarks, page 7, lines 17 to 27). Applicant argues that "under MPEP Section 2106.04(a)(2)(III)(A) claims cannot be considered to recite a mental process if they contain limitations that cannot practically be performed in the human mind, and the human mind is not equipped to perform the method of claim 1"; that "claim 1 requires 'training' a machine learning algorithm based on 'training data,' and 'iteratively adapting' the component over a 'plurality of generations' by modifying its configuration based on a verification test"; and that "processing training data to train an AI model, and iteratively recalculating fitness across multiple generations of multidimensional configurations, as claimed in claim 1 requires computational power and memory capacity that cannot practically be performed in the human mind, even with the aid of pen and paper."
Examiner respectfully disagrees. This argument is not persuasive. It is directed to a grouping of abstract ideas that the rejection does not invoke, and it proceeds from a reading of claim 1 that the claim does not bear. Each ground is independent of the other, and either alone disposes of the argument. The judicial exception identified in the Office action mailed 5/6/2026 is a mathematical concept, not a mental process. At Step 2A Prong One of that action, the first group of limitations of claim 1 is found to "recite a mathematical relationship by further organizing information and manipulating information ... through mathematical correlations"; the two further groups of limitations of claim 1, and the second group of limitations of claim 2, are each found to further "the mathematical relationship" in the same way, and claim 2's first group of limitations is found to "recite a mathematical relationship of organizing information and manipulating information ... through mathematical correlations"; and each of claims 3, 4, 5, 6 and 7 is found to recite "the same judicial exception as in claim 1". A mathematical relationship is one of the three concepts comprising the mathematical concepts grouping, which MPEP 2106.04(a)(2), subsection I, defines as "mathematical relationships, mathematical formulas or equations, and mathematical calculations", and which that subsection treats at point A. That is the grouping on which claims 1-7 stand rejected, it is unchanged in this action, and claims 1-7 remain rejected on it. The words "mental process" do not appear anywhere in that action. MPEP 2106.04(a)(2), subsection III, point A, states when a claim does not recite a mental process; it does not bear on a rejection resting on the mathematical concepts grouping, and the argument built upon it is accordingly directed to a rejection that has not been made.
The argument also overstates claim 1 in four respects, each of which is necessary to it. The four answers that follow are given on the claim as recited and hold whichever grouping of abstract ideas the argument is addressed to.
• Claim 1 does not require "training" a machine learning algorithm and recites no act of "processing training data". The recitation "the machine learning algorithm being trained based on training data to model the function" is a participial clause within the object of "defining", as set out in the preliminary matter above. The method must define a component characterized as trained; it need not train anything, hold training data, or process training data. Nothing in this recitation imposes any requirement of computational power or memory capacity on the method claim 1 recites.
• Claim 1 recites no act of "recalculating fitness". "[A] determination of a fitness" is a noun, and the clause containing it is passive and recites no actor. No act of determining or calculating fitness is recited.
• Claim 1 does not recite "multidimensional configurations". It recites "a configuration", singular and unqualified. The multidimensionality Applicant invokes is described in the Background of the disclosure (Substitute Specification, page 1, lines 14 to 22), which Applicant cites as [0003]; it is not a recitation of any claim. Claim 3's list is permissive and, being a recitation of a dependent claim, does not narrow claim 1.
• Claim 1 recites no requirement of "computational power and memory capacity". It recites no quantity, scale, rate, data volume, dimensionality, population size, resource limitation, or number of generations beyond two.
At its broadest reasonable scope, therefore, claim 1 is satisfied by defining one verification test based on two input/output pairs, defining one machine learning component having one configuration parameter, and carrying out two iterations in each of which the modification applied is taken from two candidates on a fit or not fit verdict returned by that test. Nothing in claim 1 forecloses that scope. The computational power and memory capacity Applicant invokes are attributes of an implementation the claim does not recite. This is a statement of the claim's broadest reasonable scope; it is not a finding that claim 1 recites a mental process, a grouping this rejection does not invoke.
The section Applicant quotes is also quoted in part. Although the point is not reached, the rejection resting on the mathematical concepts grouping, the same subsection states: "Nor do the courts distinguish between claims that recite mental processes performed by humans and claims that recite mental processes performed on a computer." (MPEP 2106.04(a)(2), subsection III, discussed further at point C.) See Versata Dev. Group v. SAP Am., Inc., 793 F.3d 1306, 1335 (Fed. Cir. 2015); Intellectual Ventures I LLC v. Symantec Corp., 838 F.3d 1307, 1318 (Fed. Cir. 2016); and Mortgage Grader, Inc. v. First Choice Loan Servs. Inc., 811 F.3d 1314, 1324 (Fed. Cir. 2016), which held a computer implemented method for "anonymous loan shopping" to be an abstract idea because it could be "performed by humans without a computer", the decision Applicant itself invokes at page 7, line 15.
(4) The argument that claims 2-7 are eligible because each depends from claim 1 (Remarks, page 7, lines 28 to 31). Applicant argues that "[e]ach of claim 2-7 depends from claim 1 and therefore also is not directed to an abstract idea without significantly more."
Examiner respectfully disagrees. This argument is not persuasive, and it is inaccurate as to two of the claims. Claims 6 and 7 are not dependent claims. Claim 6 is a computer system and claim 7 is a non-transitory computer-readable storage; each was analyzed in its own statutory category at Step 1 of the Office action mailed May 6, 2026. As to claims 2-5, dependency does not confer eligibility. Each claim is analyzed on the recitations it contains, and the Office action mailed 5/6/2026 analyzed claims 2, 3, 4, 5, 6 and 7 separately and in terms. Applicant engages none of those analyses and identifies no recitation of any of claims 2-7 said to render that claim eligible. The argument is a general allegation that the claims define a patentable invention without specifically pointing out how the language of the claims patentably distinguishes them. See 37 CFR 1.111(b). Claim 6 adds to the method of claim 1 only "[a] computer system comprising a processor and memory storing computer program code for performing the method of claim 1". Claim 7, which is examined as filed, adds only "[a] non-transitory computer-readable storage comprising computer program code to, when loaded into a computer system and executed thereon, cause the computer system to perform the method as claimed in claim 1". The recitations so added, a computer system, a processor, a memory, computer program code, a storage, and the loading and execution of that code on the computer system, are each recited at a high level of generality and amount to no more than generic computer components performing their ordinary functions, which is the holding of the Office action mailed 5/6/2026 as to each of claims 6 and 7. See MPEP 2106.05(f). Applicant makes no argument under Step 2B. The Step 2B holding for each of claims 1-7 is set out in the rejection above, including, for each of claims 1, 3, 4 and 5, the alternative holding that the element the Office action mailed 5/6/2026 identified as that claim's additional element does not provide an inventive concept or otherwise amount to significantly more than the abstract idea itself.
Regarding the arguments directed to the rejection of claims 1-7 under 35 U.S.C. 102(a)(1) over McDonnell.
Applicant's arguments filed 7/10/2026 with respect to the rejection of claims 1-7 under 35 U.S.C. 102(a)(1) as anticipated by McDonnell et al., US 2020/0175378 A1, have been fully considered but they are not persuasive. The rejection is maintained.
(1) The argument that McDonnell's modifications are random rather than selected from a set of candidate adaptations (Remarks, page 8, lines 9 to 10 and lines 17 to 18, and page 9, lines 1 to 7). Applicant argues that "McDonnell does not teach 'each and every element' of the claims, 'in as complete detail' as is recited in the claims"; that "McDonnell is clear in paragraphs [0074] that its mutations are random rather than being selected from a set of candidate adaptations"; that "[t]he same is true for the crossover operation 460 discussed in paragraphs [0071]-[0073]", quoting [0073]; and that "[t]herefore, McDonnell does not disclose that the adaptation for a generation is selected from a set of candidate adaptations based on a determination of a fitness of the machine learning component so adapted, as recited in claim 1."
Applicant's quotations of McDonnell are accurate. McDonnell does describe the mutation operation 470 and the crossover operation 460 as random or pseudo-random, at [0073], [0074] and [0075]. The argument is nevertheless not persuasive, for four reasons.
• First, the conclusion does not follow from the premise. Whether a modification is produced at random and whether the modification that carries into a generation is selected on fitness are two different questions. Paragraphs [0073] to [0075] answer the first. The limitation Applicant traverses addresses the second: it requires that "the adaptation for a generation is selected from a set of candidate adaptations based on a determination of a fitness". That McDonnell's operators act randomly in producing candidates says nothing about what governs which candidates carry forward.
• Second, the premise that a randomly produced modification cannot be a candidate adaptation is refuted by Applicant's own claim and Applicant's own disclosure. Claim 5, which depends from claim 1, recites "wherein at least a subset of the set of candidate adaptations is generated randomly". The disclosure provides that "[o]ne or more of the candidate adaptation 206 can be generated by, for example, a random selection of configuration parameters and/or a random adjustment to configuration parameters" (Substitute Specification, page 5, lines 4 to 6). On the reading Applicant advances, claim 5 would exclude the embodiment the specification discloses, and the disclosure would not support the claim. Randomness in the production of a candidate and fitness-based selection among candidates are not alternatives; Applicant's own disclosure combines them, permitting ties among equally fit candidates to be broken by random selection (page 5, lines 22 to 24).
• Third, the argument depends on a requirement claim 1 does not contain. The clause is recited as a passive "wherein" clause with no actor, no gerund and no temporal operator. The phrase "so adapted" fixes what the fitness is measured on, the component as adapted, and does not fix when the measurement occurs. Applicant's argument succeeds only if claim 1 required each candidate to be scored before any one of them is chosen, and claim 1 imposes no such order. The evaluate then select ordering appears in an embodiment of the disclosure, at Substitute Specification page 5, lines 12 to 15, and limitations appearing only in the specification are not read into the claims. See MPEP 2111 and 2111.01; In re Van Geuns, 988 F.2d 1181 (Fed. Cir. 1993). A scheme in which adapted candidates are produced first, each is then tested, and those found fit are carried into the succeeding generation is within the scope of the clause as recited.
• Fourth, McDonnell expressly operates on fitness criteria, in paragraphs cited in the Office action mailed May 6, 2026. McDonnell states that "[t]he fitness function 440 is an objective function that can be used to compare the models of the input set 420", and that "[i]n some examples, the fitness function 440 is based on a frequency and/or magnitude of errors produced by testing a model on the input data set 402" [0053]. "Each model of the input set 420 may be tested based on the input data set 402 to determine model fitness", and "[t]he output data of each model may be evaluated using the fitness function 440 to determine how well the model modeled the input data set 402" [0066]. "The fittest models of each 'elite species' may be identified. The fittest models overall may also be identified." [0069]. "The output set 430 may include each of the 'overall elite' models and each of the 'elite member' models. Propagating the 'overall elite' and 'elite member' models to the next epoch may preserve the 'genetic traits' resulted in caused such models being assigned high fitness values." [0070] (the quotation reproduces a printing error appearing on the face of the publication). McDonnell further recites selecting "based on a fitness function ... a subset of models from the plurality of models" on which a genetic operation is then performed [0088], and, on termination, that "an overall fittest model of the last executed epoch may be selected and output" [0076].
The sentence at [0071] on which the argument ultimately depends is "The rest of the output set 430 may be filled out by random reproduction using the crossover operation 460 and/or the mutation operation 470." "The rest" is the remainder of the output set after the elite, high fitness models of [0070] have been placed in it. The sentence therefore presupposes the fitness based selection it is offered to disprove. [0071] falls within the range "[0071]-[0073]" that Applicant itself identifies at page 9, lines 1 to 2, although Applicant quotes only [0073] from that range.
Read with the second sentence of [0071] "After the output set 430 is generated, the output set 430 may be provided as the input set 420 for the next epoch of the genetic algorithm 410" McDonnell carries out the sequence claim 1 recites, in an order claim 1 permits. The crossover operation 460 and the mutation operation 470 produce adapted models and place them in the output set 430 of an epoch [0071] to [0075]. That output set becomes the input set 420 of the next epoch [0071]. Every model of that input set is then tested with the fitness function 440 to determine its fitness [0066], [0053]. The fittest models are identified [0069]. And those models are propagated into the succeeding output set expressly because they were "assigned high fitness values" [0070]. The adaptation that carries forward into a generation is accordingly an adaptation selected out of a set of candidate adaptations on a determination of the fitness of the machine learning component as adapted, that fitness being supplied by the same fitness function 440 that constitutes the recited verification test. The limitation is met.
The traverse identifies a deficiency in one limitation of claim 1 only. Applicant states generally, at page 8, lines 9 to 10, that "McDonnell does not teach 'each and every element' of the claims, 'in as complete detail' as is recited in the claims", and at page 8, lines 11 to 16 quotes the whole of the "iteratively adapting" clause of claim 1 and asserts that it "is not disclosed or even suggested by McDonnell". The argument Applicant then makes is directed to one clause of that quotation only, that "the adaptation for a generation is selected from a set of candidate adaptations based on a determination of a fitness of the machine learning component so adapted", and it is that clause, and only that clause, that Applicant identifies at page 9, lines 5 to 7 as not disclosed. Applicant advances no argument, and identifies no deficiency in the mapping, as to the preamble, "defining a verification test for verifying that the AI algorithm models the function", "the verification test being defined based on a set of input/output pairs each indicating a required output of the function for an input", "defining a machine learning component having a machine learning algorithm and a configuration", "the machine learning algorithm being trained based on training data to model the function", "iteratively adapting the machine learning component over a plurality of generations", "wherein each generation of the machine learning component is adapted by modifying the configuration of the machine learning component", or the closing stopping-condition clause. A general allegation that a reference does not teach each and every element, unaccompanied by a showing of how the language of the claim patentably distinguishes it from the reference, does not comply with 37 CFR 1.111(b). The disclosure of each of those limitations is set out in the rejection above, and each is maintained.
(2) The argument directed to the dependent claims (Remarks, page 9, lines 10 to 13). Applicant argues that "[e]ach of the dependent claims depends from one of the independent claims referenced in the arguments presented above. Each of these claims includes features recited in its base claim as well as any intervening claims. Accordingly, these claims are allowable for substantially similar reasons as discussed above and for their additional novel features recited therein."
Examiner respectfully disagrees. This argument is not persuasive. So far as it rests on "substantially similar reasons", it is derivative of the argument directed to claim 1 and fails with it. So far as it rests on "their additional novel features recited therein", it names no feature. Applicant's arguments fail to comply with 37 CFR 1.111(b) because they amount to a general allegation that the claims define a patentable invention without specifically pointing out how the language of the claims patentably distinguishes them from the references. Claims 2, 3, 4 and 5 are mapped limitation by limitation in the rejection above, as are claims 6 and 7, and Applicant engages none of those mappings. Claims 6 and 7 are in addition not dependent claims.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KUANG FU CHEN whose telephone number is (571)272-1393. The examiner can normally be reached M-F 9:00-5:30pm ET.
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 on (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.
/KC CHEN/Primary Patent Examiner, Art Unit 2143
