INTELLIGENT SYSTEMATIC AGENT: AN ENSEMBLE OF DEEP LEARNING AND EVOLUTIONARY STRATEGIES

Application No. 18/174,000

• Filed: Feb 24, 2023
• Priority: Jul 18, 2022 — provisional 63/390,237
• Examiner: SUSSMAN MOSS, JACOB ZACHARY
• Art Unit: 2122
• Tech Center: 2100 — Computer Architecture & Software
• Assignee: Pwc Product Sales LLC
• OA Round: 2 (Final)

Office Action

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 .
This action is in response to amendments filed May 26th, 2026, in which claims 1, 20 and 21 have been amended. No claims have been cancelled nor added. The amendments have been entered, and claims 1-21 are currently pending in the case. Claims 1, 20 and 21 are independent claims.

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 1-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.

Regarding claim 1:
Step 1: Claim 1 is directed to A method, therefore it falls under the statuary category of a process.
Step 2A Prong 1: The claim recites, in part:
“generating data representing a plurality of historical episodes, wherein each historical episode is divided into a sequence of time units, wherein historical information is associated with each time unit” this encompasses the mental creation of data representing observed historical episodes divided into time units containing historical information.
“generating…for each historical episode of the plurality of episodes, a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode” this encompasses the mental creation of a respective training sequence for observed historical episodes.
“extracting a plurality of actions from one or more training action sequences” this encompasses the mental extraction of observed actions from an observed training action sequence.
“synthesizing a training data set from the extracted plurality of actions of the one or more training action sequences” this encompasses the mental creation of a training data set containing data points extracted from an observed training action sequence.
“generating…a future action for a current or future time unit” this encompasses the mental creation of a future action for a current or future time unit.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “using an evolutionary algorithm”, “generated by the evolutionary algorithm”, “using the trained deep learning model”, “training a deep learning model using the training data set to generate future actions to be executed at current or future time units” the limitation is an additional element that amounts to adding the words “apply it” (or an equivalent) with the judicial exception, or merely uses a computer in its ordinary capacity as a tool to perform an existing process. See MPEP § 2106.05(f)(2).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 2, the rejection of claim 1 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“dividing the historical information into a first information subset associated with a first set of time units and a second information subset associated with a second set of time units, wherein the first set of time units and the second set of time units are consecutive” this encompasses the mental division of observed historical information into different subsets.
“determining a scale factor based on the first information subset” this encompasses the mental determination of a scale factor based on an observed subset.
“scaling one or more values in the second information subset by the scale factor” this limitation is a mathematical concept.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “receiving the historical information”, “outputting the second set of time units and the scaled second information subset as the historical episode” these limitations are an additional element that amounts to adding insignificant extra-solution activity to the judicial exception. See MPEP § 2106.05(g).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Further, “receiving the historical information”, “outputting the second set of time units and the scaled second information subset as the historical episode” these limitations are an additional element that amounts to adding insignificant extra-solution activity to the judicial exception. See MPEP § 2106.05(g). Furthermore the additional element is directed to storing and retrieving information in memory, Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015). See MPEP § 2106.05(d)/(II). Therefore, the claim is ineligible.

Regarding claim 3, the rejection of claim 1 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“randomly generating a set of candidate action sequences corresponding to the sequence of time units for the historical episode” this encompasses the mental creation of random candidate action sequences.
“determining a set of fitness values, wherein each fitness value in the set of fitness values corresponds to a candidate action sequence in the set of candidate action sequences” this encompasses the mental determination of a set of fitness values based on an observed candidate action sequence.
“identifying, based on the set of fitness values, a fittest subset of the set of candidate action sequences” this encompasses the mental identification of a fittest subset of the set of candidate action sequences based on an observed set of fitness values.
“generating an updated set of candidate action sequences by modifying candidate action sequences in the fittest subset” this encompasses the mental creation of an updated set of candidate action sequences by modifying candidate action sequences in an observed fittest subset.
“iteratively repeating the steps of determining a set of fitness values, identifying a fittest subset, and generating an updated set of candidate action sequences” this encompasses the mental repetition of previous mental processes.
“identifying, based on the iterative repeating process, a fittest candidate action sequence” this encompasses the mental identification of an observed fittest candidate after repeating previous mental processes.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows:
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 4, the rejection of claim 3 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“the training action sequence for each historical episode is the fittest candidate action sequence corresponding to said historical episode that is identified” this encompasses the mental identification of the observed fittest candidate action sequence.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “by the evolutionary algorithm” the limitation is an additional element that amounts to adding the words “apply it” (or an equivalent) with the judicial exception, or merely uses a computer in its ordinary capacity as a tool to perform an existing process. See MPEP § 2106.05(f)(2).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 5, the rejection of claim 3 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“the plurality of cessation conditions comprise: a total number of iterations exceeds a threshold number of iterations, and one or more fitness values in the set of fitness values exceeds a threshold fitness value” this encompasses the mental repetition of mental processes until a set number of repetitions is achieves, or an observed fitness value exceeds a threshold value.
Step 2A Prong 2: The claim does not recite any additional limitations, thus does not further recite any additional elements that integrates the judicial exception into a practical application or amount to significantly more.

Regarding claim 6, the rejection of claim 3 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“modifying candidate actions sequences in the fittest subset comprises switching one or more actions in each action sequence of the fittest subset of from a first action type to a second action type” this encompasses the mental modification of an observed candidate actions sequence by switching observed actions.
Step 2A Prong 2: The claim does not recite any additional limitations, thus does not further recite any additional elements that integrates the judicial exception into a practical application or amount to significantly more.

Regarding claim 7, the rejection of claim 3 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“modifying candidate action sequences in the fittest subset of candidate action sequences comprises: selecting a first set of actions from a first action sequence of the fittest subset; selecting a second set of actions from a second action sequence of the fittest subset; and combining the first set of actions and the second set of actions to form a third action sequence” this encompasses the mental modification of an observed candidate action sequence by combining a first and second action selected based on an observed fittest subset.
Step 2A Prong 2: The claim does not recite any additional limitations, thus does not further recite any additional elements that integrates the judicial exception into a practical application or amount to significantly more.

Regarding claim 8, the rejection of claim 1 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“the historical information associated with each time unit comprises a numerical value” this encompasses the mental association of numerical values with observed historical information.
Step 2A Prong 2: The claim does not recite any additional limitations, thus does not further recite any additional elements that integrates the judicial exception into a practical application or amount to significantly more.

Regarding claim 9, the rejection of claim 8 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“each training data point of the plurality of training data points in the training data set further comprises an average value of the numerical value over a set of time units preceding a time unit in a historical episode of the plurality of historical episodes that corresponds to an action sequence from which the action in the training data point was extracted” this limitation is a mathematical concept.
Step 2A Prong 2: The claim does not recite any additional limitations, thus does not further recite any additional elements that integrates the judicial exception into a practical application or amount to significantly more.

Regarding claim 10, the rejection of claim 1 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“generating a predicted action sequence” this encompasses the mental creation of a predicted action sequence.
“comparing the predicted action sequence to the training action sequence that corresponds to the historical episode” this encompasses the mental comparison of a predicted action sequence with an observed training action sequence.
“adjusting one or more parameters…based on the comparison between the predicted action sequence and the training action sequence in the training data” this encompasses the mental adjusting of parameters based on a comparison between a predicted action sequence and an observed training action sequence
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “of the deep learning model” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 11, the rejection of claim 1 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“the future action…is configured to maximize a reward for an entity for the current or future time unit” this limitation is a mathematical concept.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “generated by the trained deep learning model” The limitation is an additional element that amounts to adding the words “apply it” (or an equivalent) with the judicial exception, or merely uses a computer in its ordinary capacity as a tool to perform an existing process. See MPEP § 2106.05(f)(2).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 12, the rejection of claim 1 is incorporated and further:
Step 2A Prong 1: a continuation of the abstract idea identified in the parent claim.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “the historical information comprises market performance information” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 13, the rejection of claim 12 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“each training action sequence…comprises, for each time unit in the historical episode associated with the training action sequence, an indication of whether to execute a purchase of an ETF at that time unit” this encompasses the mental decision of whether to execute a purchase of the ETF at a time point.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “generated by the evolutionary algorithm” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 14, the rejection of claim 13 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“the future action for the current or future time unit…comprises an indication of whether to execute a purchase of the ETF at said time unit” this encompasses the mental decision of whether to execute a purchase of the ETF at a current or future time point.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “that is generated by the deep learning model” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 15, the rejection of claim 1 is incorporated and further:
Step 2A Prong 1: a continuation of the abstract idea identified in the parent claim.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “the evolutionary algorithm is a genetic algorithm” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 16, the rejection of claim 1 is incorporated and further:
Step 2A Prong 1: a continuation of the abstract idea identified in the parent claim.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “the deep learning model is a neural network” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 17, the rejection of claim 16 is incorporated and further:
Step 2A Prong 1: a continuation of the abstract idea identified in the parent claim.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “the deep learning model is a feed-forward neural network” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 18, the rejection of claim 17 is incorporated and further:
Step 2A Prong 1: a continuation of the abstract idea identified in the parent claim.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “the feed-forward neural network comprises at least six layers” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 19, the rejection of claim 17 is incorporated and further:
Step 2A Prong 1: The claim recites, in part:
“utilizes a rectified linear unit (ReLU) activation function at one or more layers” this limitation is a mathematical concept.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “the feed-forward neural network” the limitation is an additional element that generally links the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 20:
Step 1: Claim 20 is directed to A system, therefore it falls under the statuary category of a machine.
Step 2A Prong 1: The claim recites, in part:
“generate data representing a plurality of historical episodes, wherein each historical episode is divided into a sequence of time units, wherein historical information is associated with each time unit” this encompasses the mental creation of data representing observed historical episodes divided into time units containing historical information.
“generate…for each historical episode of the plurality of episodes, a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode” this encompasses the mental creation of a respective training sequence for observed historical episodes.
“extract a plurality of actions from one or more training action sequences” this encompasses the mental extraction of observed actions from an observed training action sequence.
“synthesize a training data set from the extracted plurality of actions of the one or more training action sequences” this encompasses the mental creation of a training data set containing data points extracted from an observed training action sequence.
“generate…a future action for a current or future time unit” this encompasses the mental creation of a future action for a current or future time unit.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “using an evolutionary algorithm”, “generated by the evolutionary algorithm”, “using the trained deep learning model”, “training a deep learning model using the training data set to generate future actions to be executed at current or future time units”, “output, using the user interface, the future action to a user” the limitation is an additional element that amounts to adding the words “apply it” (or an equivalent) with the judicial exception, or merely uses a computer in its ordinary capacity as a tool to perform an existing process. See MPEP § 2106.05(f)(2).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Regarding claim 21:
Step 1: Claim 21 is directed to A non-transitory computer readable storage medium, therefore it falls under the statuary category of a manufacture.
Step 2A Prong 1: The claim recites, in part:
“generate data representing a plurality of historical episodes, wherein each historical episode is divided into a sequence of time units, wherein historical information is associated with each time unit” this encompasses the mental creation of data representing observed historical episodes divided into time units containing historical information.
“generate…for each historical episode of the plurality of episodes, a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode” this encompasses the mental creation of a respective training sequence for observed historical episodes.
“extract a plurality of actions from one or more training action sequences” this encompasses the mental extraction of observed actions from an observed training action sequence.
“synthesize a training data set from the extracted plurality of actions of the one or more training action sequences” this encompasses the mental creation of a training data set containing data points extracted from an observed training action sequence.
“generate…a future action for a current or future time unit” this encompasses the mental creation of a future action for a current or future time unit.
Step 2A Prong 2: The judicial exception is not integrated into a practical application; the remaining limitations of the claim are as follows: “using an evolutionary algorithm”, “generated by the evolutionary algorithm”, “using the trained deep learning model”, “training a deep learning model using the training data set to generate future actions to be executed at current or future time units” the limitation is an additional element that amounts to adding the words “apply it” (or an equivalent) with the judicial exception, or merely uses a computer in its ordinary capacity as a tool to perform an existing process. See MPEP § 2106.05(f)(2).
Step 2B: The additional elements, taken individually and in combination, do not provide an inventive concept of significantly more than the abstract idea itself for the reasons set forth in step 2A prong 2 above. Therefore, the claim is ineligible.

Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.

Claims 1, 3-4 and 6-21 are rejected under 35 U.S.C. 103 as being unpatentable over Chen et al. (“An Effective Approach for Obtaining a Group Trading Strategy Portfolio Using Grouping Genetic Algorithm”, Chen et al., January 23, 2019), as cited in the IDS, hereinafter Chen in view of Karathanasopoulos et al. (“Ensemble Models in Forecasting Financial Markets”, March 18, 2019) hereinafter Karathanasopoulos in view of Xie et al. (“Research on Gold ETF forecasting based on LSTM”, Xie et al., March 26, 2020) hereinafter Xie in further view of Weckman et al. ("A neural network job-shop scheduler", Weckman et al., 20 January 2008) hereinafter Weckman.

Regarding claim 1:
Chen teaches A method for training a deep learning model comprising:
generating data representing a plurality of historical episodes, wherein each historical episode is divided into a sequence of time units (Chen, page 10, col 1-2, section C, ¶2 “From Table 11, we can observe that irrespective of whether one-, two- or three-year training datasets are used for training, TABLE 13. The optimized GTSPs on the sideways trend dataset. TABLE 14. The optimized GTSPs on the downtrend dataset. the proposed approach is better than the BHS in terms of returns.” Here, the years can be considered the time units), wherein historical information is associated with each time unit (Chen, page 3, col 2, ¶5 “In Fig. 1, through the given stock price series and technical indicators, it shows that three steps are used in the data preprocessing procedure to generate the m processed TSs.” Here, the stock price series can be considered the historical episodes and the stock price can be considered the historical information.);
generating, using an evolutionary algorithm (Chen, page 3, col 2, ¶3 “The main goal of the GTSPO problem is to optimize a GTSP in accordance with objective and subjective criteria given by users using evolutionary algorithms”), for each historical episode of the plurality of episodes, a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode (Chen, page 6, col 2, ¶1 “The m selected strategies are utilized to find buy and sell signals and form the initial population (Lines 3 to 4). Every chromosome is then evaluated using the fitness function, which is composed of the four factors of the portfolio return, risk, group balance and weight balance (Lines 6 to 13). The crossover, mutation, inversion and selection operations are executed on the population to generate new offspring and the next population” Here, sell and buy signals can be considered a respective training action sequence comprising a respective sequence of actions in light of the specification, ¶20 “In some embodiments of the method, each training action sequence generated by the evolutionary algorithm comprises, for each time unit in the historical episode associated with the training action sequence, an indication of whether to execute a purchase of an ETF at that time unit.”);
Chen does not teach “training a deep learning model using the training data set to generate future actions to be executed; and
generating, using the trained deep learning model, a future action”
However, Karathanasopoulos teaches training a deep learning model using the training data set (Karathanasopoulos, page 7, ¶1 “The network parameters are then estimated by fitting the training data using the above mentioned iterative procedure (backpropagation of errors). The iteration length is optimized by maximizing a fitness function in the test dataset.”) to generate future actions to be executed (Karathanasopoulos, page 11, section 4.2, ¶1 “In this section, we present the results of the proposed methodology applied to trade the S&P 500 and Nasdaq 100 exchange trade funds in the relevant out-of-sample period.” Here, the trading of the exchange trade funds can be considered a future action); and
generating, using the trained deep learning model, a future action (Karathanasopoulos, page 11, section 4.2, ¶1 “In this section, we present the results of the proposed methodology applied to trade the S&P 500 and Nasdaq 100 exchange trade funds in the relevant out-of-sample period.” Here, the trading of the exchange trade funds can be considered a future action for a current or future time unit).
Chen and Karathanasopoulos are analogous art because both references concern learning methods for trading financial assets and portfolios. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen’s trading strategy genetic algorithm to incorporate the neural network ensemble taught by Karathanasopoulos. The motivation for doing so would have been to improve final performance, as stated in Karathanasopoulos, page 13, ¶1 “All nine ensemble NN-evolutionary models have been managed to trade successfully the two ETFs. The RBF hybrid combination models present the best performance, the second performance its coming from the RNN combinations while the MLP architectures provide the third lowest performance. All the combinations show a remarkable performance proving that hybrid combinations are improving the final performance of the model and in continuation, they are making good profit returns.”.
Chen in view of Karathanasopoulos does not teach “at current or future time units;
for a current or future time unit”
However, Xie teaches at current or future time units (Xie, page 4, col 2-1, section IV, ¶4 “The data for each consecutive m days is input, and the output is k days after m days. After processing, the input is a matrix of m rows and n columns, and the result is the vector of one k row and one column. n represents the characteristic information of each day. k represents the predicted field of view, and the predicted Adj Close price for the next k days.” Here, the next k days can be considered future time units).
for a current or future time unit (Xie, page 4, col 2-1, section IV, ¶4 “The data for each consecutive m days is input, and the output is k days after m days. After processing, the input is a matrix of m rows and n columns, and the result is the vector of one k row and one column. n represents the characteristic information of each day. k represents the predicted field of view, and the predicted Adj Close price for the next k days.” Here, the next k days can be considered future time units)
Chen in view of Karathanasopoulos and Xie are analogous art because both references concern methods for forecasting financial markets. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos’s forecasting neural network to incorporate the future time units taught by Xie. The motivation for doing so would have been to predict prices more accurately, as stated in Xie, page 5, col 2, section V, ¶1 “The model is mainly researched on the accuracy of prediction, so that the price of gold ETF can be predicted more accurately.”
Chen in view of Karathanasopoulos in further view of Xie does not teach "extracting a plurality of actions from one or more training action sequences generated by the evolutionary algorithm;
synthesizing a training data set from the extracted plurality of actions of the one or more training action sequences;"
However, Weckman teaches extracting a plurality of actions from one or more training action sequences generated by the evolutionary algorithm (Weckman, page 6, col 1, ¶3 “The 1,147 optimal schedules obtained by the GA represent a total of 41,292 scheduled operations (1,147 schedules × 36 operations/schedule).” Further, Weckman, page 2, col 1, ¶3 “In these optimized sequences, each individual operation is treated as a decision which captures some problem-specific knowledge.” Here, the 41,292 operations extracted from 1,147 schedules can be considered the plurality of actions from one or more training action sequences);
synthesizing a training data set from the extracted plurality of actions of the one or more training action sequences (Weckman, page 3, col 2, ¶4 “The second task is to model the scheduling function as a machine learning problem by defining a classification scheme, which maps the optimal sequences to data patterns. The third task is to develop a NN model with suitable architecture. The NN is trained on the classification data patterns.”);
Chen in view of Karathanasopoulos in further view of Xie and Weckman are analogous art because both references concern methods for ensembled evolutionary and deep learning techniques. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos/Xie’s forecasting neural network to incorporate the generation of training data taught by Weckman. The motivation for doing so would have been to efficiently provide training material, as stated in Weckman, page 2, col 1, ¶3 “In such a formulation, the optimal solutions generated by efficient optimizers provide the desired learning material. In these optimized sequences, each individual operation is treated as a decision which captures some problem-specific knowledge. Hence, these solutions contain valuable information such as the relationship between an operation’s attributes and its position in the sequence (solution).”

Regarding claim 3:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 1, wherein generating for each historical episode of the plurality of historical episodes, a respective training action sequence comprises:
randomly generating a set of candidate action sequences corresponding to the sequence of time units for the historical episode (Chen, page 5, col 2, section C, ¶1 “In the grouping part, it first picks two individuals randomly, where one is used as the base chromosome and the other one is the insertion chromosome. Then, parts of groups in the insertion chromosome are selected and put into the base chromosome. After insertion, groups in the base chromosome are removed if they are redundant. At last, groups are merged or split until the number of groups in a chromosome is correct. To execute the crossover operation on the weight part, the two-point crossover operator is employed to generate new chromosomes in which two points will first be determined and their sub sequences will be exchanged. Note that the appropriate arrangement should be made to ensure that the numbers of ‘1’s and ‘0’s in a chromosome are correct.” Here, the “grouping part” procedure can be considered randomly generating a set of candidate action sequences);
determining a set of fitness values, wherein each fitness value in the set of fitness values corresponds to a candidate action sequence in the set of candidate action sequences (Chen, page 7, col 3, Step 2 “The fitness value of each chromosome is calculated using the following sub-steps.”);
identifying, based on the set of fitness values, a fittest subset of the set of candidate action sequences (Chen, page 8, col 1, Step 3 “The ten chromosomes derived by the previous step are selected to form the next population using the elitist selection strategy” here, the next population formed using the elitist selection strategy can be considered a fittest subset of the set of candidate action sequences;
generating an updated set of candidate action sequences by modifying candidate action sequences in the fittest subset (Chen, page 8, col 1, Step 4 “The crossover operation is utilized in this step to generate new offspring.” Here, the new offspring can be considered an updated set of candidate action);
iteratively repeating the steps of determining a set of fitness values, identifying a fittest subset, and generating an updated set of candidate action sequences; and identifying, based on the iterative repeating process, a fittest candidate action sequence (Chen, page 4, col 1, ¶2 “The evolution is repeated until the stop conditions are reached. Finally, the best GTSP that has the highest fitness value will be delivered to users for making trading plans.” Here, the repetition of the evolution can be considered iteratively repeating the steps).

Regarding claim 4:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 3, wherein the training action sequence for each historical episode is the fittest candidate action sequence corresponding to said historical episode that is identified by the evolutionary algorithm (Karathanasopoulos, pages 12-13, section 5, ¶2 “Finding the best data optimised inputs from the backtest we start feeding and testing our ensemble models.” here, the best data optimised inputs can be considered the fittest training action sequence).
Chen/Karathanasopoulos/Xie are analogous art because both references concern learning methods for trading financial assets and portfolios. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Xie’s trading strategy genetic algorithm to incorporate the fittest candidate action sequence taught by Karathanasopoulos. The motivation for doing so would have been to improve final performance, as stated in Karathanasopoulos, pages 12-13, section 5, ¶2 “Finding the best data optimised inputs from the backtest we start feeding and testing our ensemble models. All nine ensemble NN-evolutionary models have been managed to trade successfully the two ETFs. The RBF hybrid combination models present the best performance, the second performance its coming from the RNN combinations while the MLP architectures provide the third lowest performance. All the combinations show a remarkable performance proving that hybrid combinations are improving the final performance of the model and in continuation, they are making good profit returns.”.

Regarding claim 6:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 3, wherein modifying candidate actions sequences in the fittest subset comprises switching one or more actions in each action sequence of the fittest subset of from a first action type to a second action type (Chen, page 6, col 1, ¶2 “For the mutation operations, they are performed on the trading strategy and the weight parts. To execute the mutation operation on the trading strategy part, it first chooses two groups, Gi and Gj , that both contain more than one trading strategy. Then, a trading strategy in group Gi is picked and reassigned randomly to the group Gj . With respect to the mutation operator on the weight part, two genes are selected and exchanged if they have different values. Finally, the inversion operation is executed only on the grouping part. Since the purpose of inversion operation is to increase the diversity of chromosomes, this operation exchanges two groups from the two selected groups.”).

Regarding claim 7:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 3, wherein modifying candidate action sequences in the fittest subset of candidate action sequences comprises:
selecting a first set of actions from a first action sequence of the fittest subset; selecting a second set of actions from a second action sequence of the fittest subset (Chen, Chen, page 8, col 1, Step 3 “The ten chromosomes derived by the previous step are selected to form the next population using the elitist selection strategy.”); and
combining the first set of actions and the second set of actions to form a third action sequence (Chen, page 8, col 1, Step 4 “Step 4: The crossover operation is utilized in this step to generate new offspring” here the crossover can be considered a combination from Chen, page 5, col 2, section C, ¶1 “To execute the crossover operation on the weight part, the two-point crossover operator is employed to generate new chromosomes in which two points will first be determined and their sub sequences will be exchanged.”).

Regarding claim 8:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 1, wherein the historical information associated with each time unit comprises a numerical value (Chen, page 3, col 2, ¶5 “In Fig. 1, through the given stock price series and technical indicators, it shows that three steps are used in the data preprocessing procedure to generate the m processed TSs.” Here, the stock price is a numerical value.).

Regarding claim 9:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 8, wherein each training data point of the plurality of training data points in the training data set further comprises an average value of the numerical value over a set of time units preceding a time unit in a historical episode of the plurality of historical episodes that corresponds to an action sequence from which the action in the training data point was extracted (Chen, page 9, col 2, ¶2 “The ten technical indicators are the Moving Average (MA)…” Here, the moving average can be considered an average value of the numerical value over a set of time units preceding a time unit).

Regarding claim 10:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 1, wherein training the deep learning model comprises, for each historical episode:
generating a predicted action sequence; comparing the predicted action sequence to the training action sequence that corresponds to the historical episode (Karathanasopoulos, page 7, ¶1 “Finally, the predictive value of the model is evaluated by applying it to the validation dataset (out-of-sample dataset)” here, the predictive value can be considered a predicted action sequence and the evaluation of a predictive value by applying it to the validation dataset can be considered a comparison of a predicted action with a corresponding historical episode);
adjusting one or more parameters of the deep learning model based on the comparison between the predicted action sequence and the training action sequence in the training data (Karathanasopoulos, page 6-7, section 3.2.2, ¶3 “The training of the network (which is the adjustment of its weights in the way that the network maps the input value of the training data to the corresponding output value) starts with randomly initialized weights and proceeds by applying a learning algorithm called backpropagation of errors1 (Shapiro, 2000).”).
Chen/Karathanasopoulos/Xie are analogous art because both references concern learning methods for trading financial assets and portfolios. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Xie’s trading strategy genetic algorithm to incorporate the neural network training methods taught by Karathanasopoulos. The motivation for doing so would have been to improve final performance, as stated in Karathanasopoulos, page 13, ¶1 “All nine ensemble NN-evolutionary models have been managed to trade successfully the two ETFs. The RBF hybrid combination models present the best performance, the second performance its coming from the RNN combinations while the MLP architectures provide the third lowest performance. All the combinations show a remarkable performance proving that hybrid combinations are improving the final performance of the model and in continuation, they are making good profit returns.”.

Regarding claim 11:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 1, wherein the future action generated by the trained deep learning model is configured to maximize a reward for an entity for the current or future time unit (Chen, page 3, col 2, ¶3 “For instance, the criteria could be minimizing the risk while maximizing the profit based on the allocated capitals of groups.” Here, maximizing the profit can be considered the maximizing a reward).

Regarding claim 12:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 1, wherein the historical information comprises market performance information (Chen, page 3, col 2, ¶5 “In Fig. 1, through the given stock price series and technical indicators, it shows that three steps are used in the data preprocessing procedure to generate the m processed TSs.” Here, the stock price series can be considered market performance information.).

Regarding claim 13:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 12, wherein each training action sequence generated by the evolutionary algorithm comprises, for each time unit in the historical episode associated with the training action sequence, an indication of whether to execute a purchase of an ETF at that time unit (Chen, page 6, col 2, ¶1 “The m selected strategies are utilized to find buy and sell signals and form the initial population (Lines 3 to 4). Every chromosome is then evaluated using the fitness function, which is composed of the four factors of the portfolio return, risk, group balance and weight balance (Lines 6 to 13). The crossover, mutation, inversion and selection operations are executed on the population to generate new offspring and the next population” here, the buy and sell signals can be considered the indication of whether to execute a purchase).

Regarding claim 14:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 13, wherein the future action for the current or future time unit that is generated by the deep learning model comprises an indication of whether to execute a purchase of the ETF at said time unit (Chen, page 6, col 2, ¶1 “The m selected strategies are utilized to find buy and sell signals and form the initial population (Lines 3 to 4). Every chromosome is then evaluated using the fitness function, which is composed of the four factors of the portfolio return, risk, group balance and weight balance (Lines 6 to 13). The crossover, mutation, inversion and selection operations are executed on the population to generate new offspring and the next population” here, the buy and sell signals can be considered the indication of whether to execute a purchase).

Regarding claim 15:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 1, wherein the evolutionary algorithm is a genetic algorithm (Chen, page 1, abstract “Then, an algorithm that utilizes the grouping genetic algorithm is designed for solving the GTSP optimization problem”).

Regarding claim 16:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 1, wherein the deep learning model is a neural network (Karathanasopoulos, page 6, section 3.2.2, ¶3 “The most popular and simple NN is the Multi-Layer Perceptron (MLP).”).
Chen/Karathanasopoulos/Xie are analogous art because both references concern learning methods for trading financial assets and portfolios. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Xie’s trading strategy genetic algorithm to incorporate the neural network taught by Karathanasopoulos. The motivation for doing so would have been to improve final performance, as stated in Karathanasopoulos, page 13, ¶1 “All nine ensemble NN-evolutionary models have been managed to trade successfully the two ETFs. The RBF hybrid combination models present the best performance, the second performance its coming from the RNN combinations while the MLP architectures provide the third lowest performance. All the combinations show a remarkable performance proving that hybrid combinations are improving the final performance of the model and in continuation, they are making good profit returns.”.

Regarding claim 17:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 16, wherein the deep learning model is a feed-forward neural network (Karathanasopoulos, section 3.2.2, ¶1 “Normally, each node of one layer has connections to all the other nodes of the next layer. The network processes information as follows: the input nodes contain the values of the explanatory variables. Since each node connection represents a weight factor, the information reaches a single hidden layer node as the weighted sum of its inputs. Each node of the hidden layer passes the information through a nonlinear activation function and passes it on to the output layer if the calculated value is above a threshold.”).
Chen/Karathanasopoulos/Xie are analogous art because both references concern learning methods for trading financial assets and portfolios. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Xie’s trading strategy genetic algorithm to incorporate the feed-forward neural network taught by Karathanasopoulos. The motivation for doing so would have been to improve final performance, as stated in Karathanasopoulos, page 13, ¶1 “All nine ensemble NN-evolutionary models have been managed to trade successfully the two ETFs. The RBF hybrid combination models present the best performance, the second performance its coming from the RNN combinations while the MLP architectures provide the third lowest performance. All the combinations show a remarkable performance proving that hybrid combinations are improving the final performance of the model and in continuation, they are making good profit returns.”.

Regarding claim 18:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 17, wherein the feed-forward neural network comprises at least six layers (Xie, page 3, col 2, ¶1 “Because we use a total of 11 layers of models, both TCN and LSTM use 11 layers. TCN uses an 11-layer structure with 25 channels per layer.”).
Chen in view of Karathanasopoulos and Xie are analogous art because both references concern methods for forecasting financial markets. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos’s forecasting neural network to incorporate the further layers taught by Xie. The motivation for doing so would have been to consider more prediction factors and capturing long-term dependencies, as stated in Xie, page 2, col 1, ¶3 “In this paper, we proposed a novel model named Dilated Convolution Long short-term memory (DCLSTM) considering more prediction factors and capturing long-term dependencies. The new model consists of several layer”.

Regarding claim 19:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 17, wherein the feed-forward neural network utilizes a rectified linear unit (ReLU) activation function at one or more layers (Xie, page 3, col 2, ¶4 “Also use ReLU as the activation function and use Glorot_uniform to initialize the weights.”).
Chen in view of Karathanasopoulos and Xie are analogous art because both references concern methods for forecasting financial markets. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos’s forecasting neural network to incorporate the ReLU activation function taught by Xie. The motivation for doing so would have been to consider more prediction factors and capturing long-term dependencies, as stated in Xie, page 2, col 1, ¶3 “In this paper, we proposed a novel model named Dilated Convolution Long short-term memory (DCLSTM) considering more prediction factors and capturing long-term dependencies.”.

Regarding claim 20:
Chen teaches A system for training a deep learning model comprising:
a user interface (Chen, page 12, col 1, section VII, ¶1 “After evolution, by using the optimized GTSP, various TSPs can be provided to traders for making trading plans” here, the location of plans provided to traders can be considered a user interface);
generate data representing a plurality of historical episodes, wherein each historical episode is divided into a sequence of time units (Chen, page 10, col 1-2, section C, ¶2 “From Table 11, we can observe that irrespective of whether one-, two- or three-year training datasets are used for training, TABLE 13. The optimized GTSPs on the sideways trend dataset. TABLE 14. The optimized GTSPs on the downtrend dataset. the proposed approach is better than the BHS in terms of returns.” Here, the years can be considered the time units), wherein historical information is associated with each time unit (Chen, page 3, col 2, ¶5 “In Fig. 1, through the given stock price series and technical indicators, it shows that three steps are used in the data preprocessing procedure to generate the m processed TSs.” Here, the stock price series can be considered the historical episodes and the stock price can be considered the historical information.);
generate, using an evolutionary algorithm (Chen, page 3, col 2, ¶3 “The main goal of the GTSPO problem is to optimize a GTSP in accordance with objective and subjective criteria given by users using evolutionary algorithms”), for each historical episode of the plurality of episodes, a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode (Chen, page 6, col 2, ¶1 “The m selected strategies are utilized to find buy and sell signals and form the initial population (Lines 3 to 4). Every chromosome is then evaluated using the fitness function, which is composed of the four factors of the portfolio return, risk, group balance and weight balance (Lines 6 to 13). The crossover, mutation, inversion and selection operations are executed on the population to generate new offspring and the next population” Here, sell and buy signals can be considered a respective training action sequence comprising a respective sequence of actions in light of the specification, ¶20 “In some embodiments of the method, each training action sequence generated by the evolutionary algorithm comprises, for each time unit in the historical episode associated with the training action sequence, an indication of whether to execute a purchase of an ETF at that time unit.”);
output, using the user interface, the future action to a user (Chen, page 4, col 1, ¶2 “Finally, the best GTSP that has the highest fitness value will be delivered to users for making trading plans.”).
Chen does not teach “training a deep learning model using the training data set to generate future actions to be executed; and
generate, using the trained deep learning model, a future action”
However, Karathanasopoulos teaches training a deep learning model using the training data set (Karathanasopoulos, page 7, ¶1 “The network parameters are then estimated by fitting the training data using the above mentioned iterative procedure (backpropagation of errors). The iteration length is optimized by maximizing a fitness function in the test dataset.”) to generate future actions to be executed (Karathanasopoulos, page 11, section 4.2, ¶1 “In this section, we present the results of the proposed methodology applied to trade the S&P 500 and Nasdaq 100 exchange trade funds in the relevant out-of-sample period.” Here, the trading of the exchange trade funds can be considered a future action); and
generate, using the trained deep learning model, a future action (Karathanasopoulos, page 11, section 4.2, ¶1 “In this section, we present the results of the proposed methodology applied to trade the S&P 500 and Nasdaq 100 exchange trade funds in the relevant out-of-sample period.” Here, the trading of the exchange trade funds can be considered a future action).
Chen and Karathanasopoulos are analogous art because both references concern learning methods for trading financial assets and portfolios. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen’s trading strategy genetic algorithm to incorporate the neural network ensemble taught by Karathanasopoulos. The motivation for doing so would have been to improve final performance, as stated in Karathanasopoulos, page 13, ¶1 “All nine ensemble NN-evolutionary models have been managed to trade successfully the two ETFs. The RBF hybrid combination models present the best performance, the second performance its coming from the RNN combinations while the MLP architectures provide the third lowest performance. All the combinations show a remarkable performance proving that hybrid combinations are improving the final performance of the model and in continuation, they are making good profit returns.”.
Chen in view of Karathanasopoulos does not teach “at current or future time units;
for a current or future time unit”
However, Xie teaches at current or future time units (Xie, page 4, col 2-1, section IV, ¶4 “The data for each consecutive m days is input, and the output is k days after m days. After processing, the input is a matrix of m rows and n columns, and the result is the vector of one k row and one column. n represents the characteristic information of each day. k represents the predicted field of view, and the predicted Adj Close price for the next k days.” Here, the next k days can be considered future time units).
for a current or future time unit (Xie, page 4, col 2-1, section IV, ¶4 “The data for each consecutive m days is input, and the output is k days after m days. After processing, the input is a matrix of m rows and n columns, and the result is the vector of one k row and one column. n represents the characteristic information of each day. k represents the predicted field of view, and the predicted Adj Close price for the next k days.” Here, the next k days can be considered future time units)
Chen in view of Karathanasopoulos and Xie are analogous art because both references concern methods for forecasting financial markets. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos’s forecasting neural network to incorporate the future time units taught by Xie. The motivation for doing so would have been to predict prices more accurately, as stated in Xie, page 5, col 2, section V, ¶1 “The model is mainly researched on the accuracy of prediction, so that the price of gold ETF can be predicted more accurately.”
Chen in view of Karathanasopoulos in further view of Xie does not teach "extract a plurality of actions from one or more training action sequences generated by the evolutionary algorithm;
synthesize a training data set from the extracted plurality of actions of the one or more training action sequences;"
However, Weckman teaches extract a plurality of actions from one or more training action sequences generated by the evolutionary algorithm (Weckman, page 6, col 1, ¶3 “The 1,147 optimal schedules obtained by the GA represent a total of 41,292 scheduled operations (1,147 schedules × 36 operations/schedule).” Further, Weckman, page 2, col 1, ¶3 “In these optimized sequences, each individual operation is treated as a decision which captures some problem-specific knowledge.” Here, the 41,292 operations extracted from 1,147 schedules can be considered the plurality of actions from one or more training action sequences);
synthesize a training data set from the extracted plurality of actions of the one or more training action sequences (Weckman, page 3, col 2, ¶4 “The second task is to model the scheduling function as a machine learning problem by defining a classification scheme, which maps the optimal sequences to data patterns. The third task is to develop a NN model with suitable architecture. The NN is trained on the classification data patterns.”);
Chen in view of Karathanasopoulos in further view of Xie and Weckman are analogous art because both references concern methods for ensembled evolutionary and deep learning techniques. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos/Xie’s forecasting neural network to incorporate the generation of training data taught by Weckman. The motivation for doing so would have been to efficiently provide training material, as stated in Weckman, page 2, col 1, ¶3 “In such a formulation, the optimal solutions generated by efficient optimizers provide the desired learning material. In these optimized sequences, each individual operation is treated as a decision which captures some problem-specific knowledge. Hence, these solutions contain valuable information such as the relationship between an operation’s attributes and its position in the sequence (solution).”

Regarding claim 21:
Chen teaches A non-transitory computer readable storage medium storing instructions that, when executed by one or more processors of an electronic device, cause the electronic device to:
generate data representing a plurality of historical episodes, wherein each historical episode is divided into a sequence of time units (Chen, page 10, col 1-2, section C, ¶2 “From Table 11, we can observe that irrespective of whether one-, two- or three-year training datasets are used for training, TABLE 13. The optimized GTSPs on the sideways trend dataset. TABLE 14. The optimized GTSPs on the downtrend dataset. the proposed approach is better than the BHS in terms of returns.” Here, the years can be considered the time units), wherein historical information is associated with each time unit (Chen, page 3, col 2, ¶5 “In Fig. 1, through the given stock price series and technical indicators, it shows that three steps are used in the data preprocessing procedure to generate the m processed TSs.” Here, the stock price series can be considered the historical episodes and the stock price can be considered the historical information.);
generate, using an evolutionary algorithm (Chen, page 3, col 2, ¶3 “The main goal of the GTSPO problem is to optimize a GTSP in accordance with objective and subjective criteria given by users using evolutionary algorithms”), for each historical episode of the plurality of episodes, a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode (Chen, page 6, col 2, ¶1 “The m selected strategies are utilized to find buy and sell signals and form the initial population (Lines 3 to 4). Every chromosome is then evaluated using the fitness function, which is composed of the four factors of the portfolio return, risk, group balance and weight balance (Lines 6 to 13). The crossover, mutation, inversion and selection operations are executed on the population to generate new offspring and the next population” Here, sell and buy signals can be considered a respective training action sequence comprising a respective sequence of actions in light of the specification, ¶20 “In some embodiments of the method, each training action sequence generated by the evolutionary algorithm comprises, for each time unit in the historical episode associated with the training action sequence, an indication of whether to execute a purchase of an ETF at that time unit.”);
Chen does not teach “training a deep learning model using the training data set to generate future actions to be executed; and
generate, using the trained deep learning model, a future action”
However, Karathanasopoulos teaches training a deep learning model using the training data set (Karathanasopoulos, page 7, ¶1 “The network parameters are then estimated by fitting the training data using the above mentioned iterative procedure (backpropagation of errors). The iteration length is optimized by maximizing a fitness function in the test dataset.”) to generate future actions to be executed (Karathanasopoulos, page 11, section 4.2, ¶1 “In this section, we present the results of the proposed methodology applied to trade the S&P 500 and Nasdaq 100 exchange trade funds in the relevant out-of-sample period.” Here, the trading of the exchange trade funds can be considered a future action); and
generate, using the trained deep learning model, a future action (Karathanasopoulos, page 11, section 4.2, ¶1 “In this section, we present the results of the proposed methodology applied to trade the S&P 500 and Nasdaq 100 exchange trade funds in the relevant out-of-sample period.” Here, the trading of the exchange trade funds can be considered a future action).
Chen and Karathanasopoulos are analogous art because both references concern learning methods for trading financial assets and portfolios. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen’s trading strategy genetic algorithm to incorporate the neural network ensemble taught by Karathanasopoulos. The motivation for doing so would have been to improve final performance, as stated in Karathanasopoulos, page 13, ¶1 “All nine ensemble NN-evolutionary models have been managed to trade successfully the two ETFs. The RBF hybrid combination models present the best performance, the second performance its coming from the RNN combinations while the MLP architectures provide the third lowest performance. All the combinations show a remarkable performance proving that hybrid combinations are improving the final performance of the model and in continuation, they are making good profit returns.”.
Chen in view of Karathanasopoulos does not teach “at current or future time units;
for a current or future time unit”
However, Xie teaches at current or future time units (Xie, page 4, col 2-1, section IV, ¶4 “The data for each consecutive m days is input, and the output is k days after m days. After processing, the input is a matrix of m rows and n columns, and the result is the vector of one k row and one column. n represents the characteristic information of each day. k represents the predicted field of view, and the predicted Adj Close price for the next k days.” Here, the next k days can be considered future time units).
for a current or future time unit (Xie, page 4, col 2-1, section IV, ¶4 “The data for each consecutive m days is input, and the output is k days after m days. After processing, the input is a matrix of m rows and n columns, and the result is the vector of one k row and one column. n represents the characteristic information of each day. k represents the predicted field of view, and the predicted Adj Close price for the next k days.” Here, the next k days can be considered future time units)
Chen in view of Karathanasopoulos and Xie are analogous art because both references concern methods for forecasting financial markets. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos’s forecasting neural network to incorporate the future time units taught by Xie. The motivation for doing so would have been to predict prices more accurately, as stated in Xie, page 5, col 2, section V, ¶1 “The model is mainly researched on the accuracy of prediction, so that the price of gold ETF can be predicted more accurately.”
Chen in view of Karathanasopoulos in further view of Xie does not teach "extract a plurality of actions from one or more training action sequences generated by the evolutionary algorithm;
synthesize a training data set from the extracted plurality of actions of the one or more training action sequences;"
However, Weckman teaches extract a plurality of actions from one or more training action sequences generated by the evolutionary algorithm (Weckman, page 6, col 1, ¶3 “The 1,147 optimal schedules obtained by the GA represent a total of 41,292 scheduled operations (1,147 schedules × 36 operations/schedule).” Further, Weckman, page 2, col 1, ¶3 “In these optimized sequences, each individual operation is treated as a decision which captures some problem-specific knowledge.” Here, the 41,292 operations extracted from 1,147 schedules can be considered the plurality of actions from one or more training action sequences);
synthesize a training data set from the extracted plurality of actions of the one or more training action sequences (Weckman, page 3, col 2, ¶4 “The second task is to model the scheduling function as a machine learning problem by defining a classification scheme, which maps the optimal sequences to data patterns. The third task is to develop a NN model with suitable architecture. The NN is trained on the classification data patterns.”);
Chen in view of Karathanasopoulos in further view of Xie and Weckman are analogous art because both references concern methods for ensembled evolutionary and deep learning techniques. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos/Xie’s forecasting neural network to incorporate the generation of training data taught by Weckman. The motivation for doing so would have been to efficiently provide training material, as stated in Weckman, page 2, col 1, ¶3 “In such a formulation, the optimal solutions generated by efficient optimizers provide the desired learning material. In these optimized sequences, each individual operation is treated as a decision which captures some problem-specific knowledge. Hence, these solutions contain valuable information such as the relationship between an operation’s attributes and its position in the sequence (solution).”

Claim 2 is rejected under 35 U.S.C. 103 as being unpatentable over Chen in view of Karathanasopoulos in view of Xie in view of Weckman in further view of Chakravorty et al. (“Deep Learning based Global Tactical Asset Allocation”, Chakravorty et al., October 19, 2018) hereinafter Chakravorty.

Regarding claim 2:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 1,
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman does not teach “wherein generating a historical episode of the plurality of historical episodes comprises:
receiving the historical information;
dividing the historical information into a first information subset associated with a first set of time units and a second information subset associated with a second set of time units, wherein the first set of time units and the second set of time units are consecutive;
determining a scale factor based on the first information subset; scaling one or more values in the second information subset by the scale factor; and
outputting the second set of time units and the scaled second information subset as the historical episode”
However, Chakravorty teaches wherein generating a historical episode of the plurality of historical episodes comprises:
receiving the historical information (Chakravorty, page 3, ¶1 “We also discuss the different features derived from end of day price-volume and macroeconomic data which are used as inputs to the deep learning models.”);
dividing the historical information into a first information subset associated with a first set of time units and a second information subset associated with a second set of time units, wherein the first set of time units and the second set of time units are consecutive (Chakravorty, page 5, ¶2 “We use historical data since 1995. Data till the end of 2013 is used for training and validation purposes and the remaining as the test data that is only used to evaluate the results once the investment strategy has been fixed” here, the data from 1995-2013 can be considered a first information subset associated with a first set of time units and the remaining can be considered a second information subset associated with a second set of time units);
determining a scale factor based on the first information subset; scaling one or more values in the second information subset by the scale factor (Chakravorty, page 5, ¶2 “Every time we retrain our model we calculate a new mean and standard deviation for our data to normalize it using a z-score (applied column-wise). We use these same means and standard deviations to normalize all new data points that come in until the next time we have to retrain our model, at which time we will calculate a new mean and standard deviation to normalize our data” here, the calculation of a new mean and standard deviation and normalizing by the z-score can be considered determining a scaling factor and scaling in light of the specification, ¶86 “In some embodiments, scaling the second thirty-day subset may comprise normalizing the second thirty-day subset of historical data based on a mean ETF price value of the first thirty-day subset, a mean ETF price value of the second thirty-day subset, a standard deviation of the ETF price values of the first thirty-day subset, and/or a standard deviation of the ETF price values of the second thirty-day subset.”); and
outputting the second set of time units and the scaled second information subset as the historical episode (Chakravorty, page 5, ¶2 “Every time we retrain our model we calculate a new mean and standard deviation for our data to normalize it using a z-score (applied column-wise).” Here, the data for retraining that has been normalized can be considered the second set of time units and the scaled second information ).
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman and Chakravorty are analogous art because both references concern methods for trading financial assets and portfolios. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos/Xie/Weckman’s financial learning system to incorporate the scaling factors taught by Chakravorty. The motivation for doing so would have been to achieve better performance as stated in Chakravorty, page 12, figure 10 “The utility based portfolio construction leads to much better performance compared to risk parity.”.

Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Chen in view of Karathanasopoulos in view of Xie in further view of Weckman in further view of Yarsky (“Using a genetic algorithm to fit parameters of a COVID-19 SEIR model for US states”, Yarsky, February, 13 2021).

Regarding claim 5:
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman teaches The method of claim 3, wherein the iteratively repeating continues until at least one cessation condition of plurality of cessation conditions is met, wherein the plurality of cessation conditions comprise:
a total number of iterations exceeds a threshold number of iterations (Chen, page 6, Figure 6, The pseudo code of the proposed approach, line 5 “For I = 0 to numGeneration Do” here, numGeneration can be considered a threshold number of iterations),
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman does not teach “one or more fitness values in the set of fitness values exceeds a threshold fitness value “
However, Yarsky teaches one or more fitness values in the set of fitness values exceeds a threshold fitness value (Yarsky, page 2, section 2, ¶6 “Successive generations are simulated until an agent produces results within an acceptable fitness threshold, or the total number of maximum generations are simulated.”).
Chen in view of Karathanasopoulos in view of Xie in further view of Weckman and Yarksy are analogous art because both references concern methods for using genetic algorithms. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Chen/Karathanasopoulos/Xie/Weckman’s genetic algorithm to incorporate the acceptable fitness threshold taught by Yarsky. The motivation for doing so would have been to have good agreement between the model and available data as stated in Yarsky, Abstract “Use of the genetic algorithm produces exceptionally good agreement between the model and available data.”.

Response to Arguments
Applicant's arguments filed May 26th, 2026 (hereinafter “Remarks”) have been fully considered but they are not persuasive.
Regarding the objections to the drawings, Applicant’s amended specification has overcome the objections, which are withdrawn.

Rejections under 35 U.S.C. § 101:
Argument 1:
“Claim 1 does not recite an abstract idea at least because it recites limitations that cannot practically be performed in the human mind.” (Remarks, page 14).

Examiners Response:
Examiner respectfully disagrees, the MPEP states “…the courts distinguish between claims that recite mental processes performed by humans and claims that recite mental processes performed on a computer. As the Federal Circuit has explained, "[c]ourts have examined claims that required the use of a computer and still found that the underlying, patent-ineligible invention could be performed via pen and paper or in a person’s mind." Versata Dev. Group v. SAP Am., Inc., 793 F.3d 1306, 1335, 115 USPQ2d 1681, 1702 (Fed. Cir. 2015). See also Intellectual Ventures I LLC v. Symantec Corp., 838 F.3d 1307, 1318, 120 USPQ2d 1353, 1360 (Fed. Cir. 2016) (‘‘[W]ith the exception of generic computer-implemented steps, there is nothing in the claims themselves that foreclose them from being performed by a human, mentally or with pen and paper.’’); Mortgage Grader, Inc. v. First Choice Loan Servs. Inc., 811 F.3d 1314, 1324, 117 USPQ2d 1693, 1699 (Fed. Cir. 2016) (holding that computer-implemented method for "anonymous loan shopping" was an abstract idea because it could be "performed by humans without a computer"). Mental processes recited in claims that require computers are explained further below with respect to point C.” See MPEP § 2106.04(a)(2)(III). Here, the organization, generation and synthesization of observed data could practically be performed via pen and paper or in a person’s mind. The use of genetic algorithms and deep learning models generally links the use of the judicial exception to a particular technological environment or field of use or amounts to adding the words “apply it” (or an equivalent) with the judicial exception, or merely uses a computer in its ordinary capacity as a tool to perform an existing process. See MPEP §§ 2106.05(h), 2106.05(f)(2).

Argument 2:
“The explanation provided for Example 39's eligibility includes that "the claim does not recite a mental process because the steps are not practically performed in the human mind." The example further indicates that although the claim involves mathematical concepts, the concepts are not recited in the claim and that the claim does not recite any method of organizing human activity. Thus "the claim is eligible because it does not recite a judicial exception." Similarly here, claim 1 does not recite any mathematical formulas or equations or any certain method of organizing human activity. Rather, like Example 39, claim 1 recites specific processing steps for training a deep learning model that cannot practically be performed in the human mind. Accordingly, claim 1 does not recite a judicial exception for similar reasons to those described with reference to Example 39's eligibility.” (Remarks, page 15).

Examiners Response:
Examiner respectfully disagrees, the MPEP states “An inventive concept "cannot be furnished by the unpatentable law of nature (or natural phenomenon or abstract idea) itself." Genetic Techs. v. Merial LLC, 818 F.3d 1369, 1376, 118 USPQ2d 1541, 1546 (Fed. Cir. 2016).” See MPEP § 2106.05(I). Further, “It is important to note, the judicial exception alone cannot provide the improvement.” See MPEP § 2106.05(a). Unlike in example 39, which does not recite a judicial exception and is thus eligible, the presently filed claims have been found to be directed to a judicial exception in Step 2A, prong 1, and are thus practically performable in the human mind (or with pen and paper).

Argument 3:
“Even assuming, arguendo, that claim 1 recites an abstract idea, amended claim 1 as a whole integrates any such concept into a practical application. A claim integrates a judicial exception into a practical application when recites an improvement to a technology or technical field. (MPEP 2106.04 II.A.2.) Improvements to machine learning constitute improvements to a technology or technical field. See Ex Parte Desjardins, Appeal No. 2024-000567 (PTAB September 26, 2025, Appeals Review Panel Decision) (precedential). Like the claims at issue in Desjardins, claim 1 recites limitations reflecting a technological improvement in machine learning.” (Remarks, page 16).

Examiners Response:
Examiner respectfully disagrees, the MPEP states “It should be noted that while this consideration is often referred to in an abbreviated manner as the "improvements consideration," the word "improvements" in the context of this consideration is limited to improvements to the functioning of a computer or any other technology/technical field, whether in Step 2A Prong Two or in Step 2B.” See MPEP § 2106.04(d)(1). Further, the additional elements are recited at a high level of generality, and even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application.

Argument 4:
““This may allow the methods to be adapted for use by a variety of entity types (ranging from individual people, who may not have access to high-powered computers, to large companies with abundant resources)”…Thus, claim 1 recites the specific processing steps by which an evolutionary algorithm generates particular training data (for each historical episode of the plurality of episodes, a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode), from which a plurality of actions are extracted, and once extracted, used to synthesize a training data set from the extracted plurality of actions of the one or more training action sequences. The training dataset is then used to train the deep learning model, thus enabling the improvements described in the specification.” (Remarks, pages 17-18).

Examiners Response:
Examiner respectfully disagrees, the MPEP states “Use of a computer or other machinery in its ordinary capacity for economic or other tasks (e.g., to receive, store, or transmit data) or simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., a fundamental economic practice or mathematical equation) does not integrate a judicial exception into a practical application or provide significantly more. See Affinity Labs v. DirecTV, 838 F.3d 1253, 1262, 120 USPQ2d 1201, 1207 (Fed. Cir. 2016) (cellular telephone); TLI Communications LLC v. AV Auto, LLC, 823 F.3d 607, 613, 118 USPQ2d 1744, 1748 (Fed. Cir. 2016) (computer server and telephone unit). Similarly, "claiming the improved speed or efficiency inherent with applying the abstract idea on a computer" does not integrate a judicial exception into a practical application or provide an inventive concept. Intellectual Ventures I LLC v. Capital One Bank (USA), 792 F.3d 1363, 1367, 115 USPQ2d 1636, 1639 (Fed. Cir. 2015). In contrast, a claim that purports to improve computer capabilities or to improve an existing technology may integrate a judicial exception into a practical application or provide significantly more. McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299, 1314-15, 120 USPQ2d 1091, 1101-02 (Fed. Cir. 2016); Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1335-36, 118 USPQ2d 1684, 1688-89 (Fed. Cir. 2016). See MPEP §§ 2106.04(d)(1) and 2106.05(a) for a discussion of improvements to the functioning of a computer or to another technology or technical field.” See MPEP § 2106.5(f). Further, “Conversely, if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology. Second, 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 claim itself does not need to explicitly recite the improvement described in the specification (e.g., "thereby increasing the bandwidth of the channel").” See MPEP § 2106.04(d)(1). Here, the use of a generic evolutionary algorithm and deep learning model on generic computing equipment does not integrate a judicial exception into a practical application or provide significantly more. The applicant merely uses a computer to perform processes which can be performed by a human mind. An improvement to creating training data may be an improvement in an abstract idea, but not an improvement in the functioning of a computer or other technology.

Rejections under 35 U.S.C. § 103:
Argument 5:
“The cited references, taken either alone or in combination, fail to disclose or suggest at least "generating, using an evolutionary algorithm, for each historical episode of the plurality of episodes, a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode," as recited in amended claim…Thus, the Examiner is mapping the buy/sell signals described in Chen to "a respective training action sequence comprising a respective sequence of actions that corresponds to the sequence of time units for the historical episode." This mapping fundamentally mischaracterizes what Chen's evolutionary algorithm does. Chen describes that the purpose of his approach is "to optimize a trading strategy portfolio, which is a set of strategies where the return and risk of the portfolio can be maximized and minimized, respectively." (Chen, Abstract.) Chen explains that "technical or fundamental indicators are employed to generate trading strategies," and gives the example: "When the CCI value of a stock crosses -100 from the bottom, then a buy signal is suggested" and "When the CCI value of a stock crosses 100 from the top, a sell signal is suggested." (Chen, p. 1, col. 2.) Importantly, the evolutionary algorithm in Chen does not generate the buy and sell signals. The signals come from the trading strategies themselves, which are predefined rules. The evolutionary algorithm in Chen only decides how to group those strategies together.” (Remarks, page 19-20).

Examiners Response:
Examiner respectfully disagrees, the MPEP states “Because applicant has the opportunity to amend the claims during prosecution, giving a claim its broadest reasonable interpretation will reduce the possibility that the claim, once issued, will be interpreted more broadly than is justified. In re Yamamoto, 740 F.2d 1569, 1571 (Fed. Cir. 1984); In re Zletz, 893 F.2d 319, 321, 13 USPQ2d 1320, 1322 (Fed. Cir. 1989) ("During patent examination the pending claims must be interpreted as broadly as their terms reasonably allow."); In re Prater, 415 F.2d 1393, 1404-05, 162 USPQ 541, 550-51 (CCPA 1969) (Claim 9 was directed to a process of analyzing data generated by mass spectrographic analysis of a gas. The process comprised selecting the data to be analyzed by subjecting the data to a mathematical manipulation. The examiner made rejections under 35 U.S.C. 101 and 35 U.S.C. 102. In the 35 U.S.C. 102 rejection, the examiner explained that the claim was anticipated by a mental process augmented by pencil and paper markings. The court agreed that the claim was not limited to using a machine to carry out the process since the claim did not explicitly set forth the machine. The court explained that "reading a claim in light of the specification, to thereby interpret limitations explicitly recited in the claim, is a quite different thing from ‘reading limitations of the specification into a claim,’ to thereby narrow the scope of the claim by implicitly adding disclosed limitations which have no express basis in the claim." The court found that applicant was advocating the latter, i.e., the impermissible importation of subject matter from the specification into the claim.).” The broadest reasonable interpretation of a respective training action sequence comprising a respective sequence of actions, as claimed, includes using a genetic algorithm to create a series of buy and sell signals. Like Chen, the broadest reasonable interpretation of the claim does not require the actions to be the chromosomes of the genetic algorithm, just that the action sequences are generated using the evolutionary algorithm. Therefore, Chen can be considered to teach the generation of the training action sequence.

Argument 6:
“The cited references, taken either alone or in combination, fail to disclose or suggest at least "extracting a plurality of actions from one or more training action sequences generated by the evolutionary algorithm; [and] synthesizing a training data set from the extracted plurality of actions of the one or more training action sequences," as recited in amended claim 1.” (Remarks, page 21).

Examiners Response:
Applicant’s arguments have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.

Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Mirjalili et al. ("Evolutionary Algorithms and Neural Networks Theory and Applications", Mirjalili et al., 2019) discloses theory and application of evolutionary algorithms and artificial neural networks. And an attempt is made to make a bridge between these two fields with an emphasis on real-world applications. Part I presents well-regarded and recent evolutionary algorithms and optimisation techniques. Quantitative and qualitative analyses of each algorithm are per formed to understand the behaviour and investigate their potentials to be used in conjunction with artificial neural networks.

Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JACOB Z SUSSMAN MOSS whose telephone number is [phone]. The examiner can normally be reached Monday - Friday, 9 a.m. - 5 p.m. 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, Kakali Chaki can be reached on [phone]. 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.S.M./Examiner, Art Unit 2122

/KAKALI CHAKI/Supervisory Patent Examiner, Art Unit 2122

Prosecution Timeline

• Feb 24, 2023: Application Filed
• Dec 23, 2025: Non-Final Rejection mailed — §101, §103
• Mar 18, 2026: Examiner Interview Summary
• Mar 18, 2026: Applicant Interview (Telephonic)
• May 26, 2026: Response Filed
• Aug 18, 2026: Final Rejection mailed — §101, §103 (current)