DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 02/24/2026 has been entered.
The status of the claims is as follows.
Claims 1, 2, 4, 16 and 20 are amended. Claims 9 and 19 have been canceled. Claims 1-8, 10-18, and 20 are currently pending.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 4, 10-12, 16, and 20 are rejected under 35 U.S.C. 103 as being anticipated by Mitchko et al. (US20220215465A1, hereinafter “Mitchko”) in view of Lopes et al. (US20210350382A1, hereinafter “Lopes”) in view of Armelin et al (US20220397425, hereinafter “Armelin”) and further in view of Kim et al. (US20220180469A1, hereinafter “Kim”)
Regarding Claim 1,
Mitchko discloses receiving, with at least one processor, payment transaction time series data associated with a plurality of payment transactions in a time series, the plurality of payment transactions including a first subset of payment transactions associated with a first entity; (Mitchko [0033]; “The one or more inputs may comprise transaction data and/or historical datasets associated with a particular user. The historical datasets may comprise a pattern of the user's purchases and/or account balances over a particular period of time. Additionally or alternatively, the one or more inputs may comprise identifying one or more users or groups of users that have similar spending habits, account balances, income, etc. After the one or more users or groups of users have been identified, transaction data and/or historical datasets of the one or more users or groups of users may be chosen as an input for the predictive model” wherein purchase transaction data of several users inputted into a model reads on receiving a plurality of payment transaction data of at least a first entity; wherein the historical datasets comprising the patterns of user purchases and spending habits over a particular period of time thus implicitly reads on a time series)
generating, with the at least one processor, at least one prediction of a net settlement position of the first entity … and a plurality of machine learning models; (Mitchko [0029]; “By monitoring the bank account levels, transaction history, incoming and outgoing expenses, and/or credit card activity of the user, the analysis application 245 may train the predictive model (e.g., machine learning model) to better forecast the user's spending habits, including the incoming funds and/or outgoing expenses associated with a user's bank account. The analysis application 245 may be able to predict the balance of the user's bank account when payment for a credit transaction will be due, for example, based on the analysis application 245 monitoring credit transactions through the transaction database 250.” wherein the machine learning model providing predictions of the user’s bank account balance for a given transaction based on user transaction history, expenses, and bank account levels reads on generating a prediction of a net settlement position of the first entity)
communicating, with the at least one processor, the at least one prediction of the net settlement position to a first entity system associated with the first entity. (Mitchko [Fig. 2];
PNG
media_image1.png
463
606
media_image1.png
Greyscale
Mitchko [0029]; “The analysis application 245 may be able to predict the balance of the user's bank account when payment for a credit transaction will be due, for example, based on the analysis application 245 monitoring credit transactions through the transaction database 250. Monitoring transaction database 250 may also help the predictive model (e.g., machine learning model) to identify fraudulent transactions. For example, if a customer usually spends less than $40 per transaction, an alert may be generated on a purchase of greater than $300. The alert may flag the purchase as a fraudulent transaction. Additionally or alternatively, the alert may trigger a user verification of the transaction. In this regard, a financial institution (e.g., a bank, a creditor, a credit card issuer, etc.) may contact the consumer to verify the transaction. If the consumer verifies the transaction, the alert may be cleared.” wherein the alert based on predicted user account balances and net transactions being sent to user associated device reads on communicating a prediction of net settlement position to a first entity system)
Mitchko fails to explicitly disclose but Lopes discloses determining, with the at least one processor, a plurality of features based on the payment transaction time series data associated with the plurality of payment transactions; (Lopes [0066]; “In some implementations, the machine learning system may perform natural language processing and/or another feature identification technique to extract features (e.g., variable types) and/or feature values (e.g., variable values) from text (e.g., unstructured data) input to the machine learning system”)
Mitchko discloses generating, with the at least one processor, at least one prediction of a net settlement position of the first entity … and a plurality of machine learning models. Mitchko does not explicitly discloses feature extraction to determine the plurality of features. However, Lopes discloses features extraction to determine a plurality of features. By performing Lopes’ feature extraction on the data used as input to Mitchko’s machine learning models for prediction generation, the combination discloses generating, with the at least one processor, at least one prediction of a net settlement position of the first entity based on the plurality of features and a plurality of machine learning models.
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko’s method of predicting net settlement positions in interactions based on transaction features to use Lopes’ method of extracting the transaction features from received data observations. The motivation to do so is because “In some cases, different observations may be associated with different sets of variable values, sometimes referred to as feature values. In some implementations, the machine learning system may determine variable values for a specific observation based on input received from one or more transaction data sources” (Lopes [0066]), thus allowing Mitchko’s model to be able to utilize input features from a wider range of observations and data types.
The combination of Mitchko/Lopes fails to explicitly disclose but Armelin discloses generating denoised features based on inputting the plurality of features to a denoising autoencoder (DAE) (Armelin [0019]; “A recurrent denoising autoencoder is a special type of autoencoder. An autoencoder is an encoder-decoder neural network that learns to copy its input to its output. It has an internal (hidden) layer that describes a code used to represent the input, and it is constituted by two main parts: an encoder that maps the input into the code, and a decoder that maps the code to a reconstruction of the original input. Denoising autoencoders (DAEs) take a partially corrupted input and are trained to recover the original undistorted input. In practice, the objective of denoising autoencoders is that of cleaning the corrupted input, or denoising. Recurrent denoising autoencoders can summarize sequential data through an encoder structure into a fixed-length vector and then reconstruct into its original sequential form through the decoder structure. The summarized information can be used to represent time series features.”
Armelin [0051]; “FIG. 2 a) illustrates an embodiment of the proposed solution to this problem by introducing a machine learning circuitry 220 to separate the useful signal component 202b from the attack signal component 202a of the manipulated measurement signal 202c in order to denoise the same. For denoising the noisy measurement signal generated by the MEMS sensor, the machine learning circuitry 220 may learn a denoising model mapping a noisy measurement signal x to a reconstructed useful signal ŷ 203 according to a model function f(x).fwdarw.ŷ … In order to mitigate manipulations of the measurements signal generated by the MEMS sensor circuit 210, the machine learning circuitry 220 may learn the model function for denoising the noisy measurement signal of an attacked MEMS sensor. Suitable models for denoising may be obtained by recurrent neural networks (RNNs) or recurrent denoising autoencoders (RDAEs). The architecture of a RDAE is based on recurrent neural networks (RNNs) and denoising autoencoders (DAEs).”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes’ method of predicting processed volume to incorporate Armelin’s recurrent denoising autoencoder for preprocessing raw data. The motivation to do so is to allow Mitchko/Lopes’ model to “clean the corrupted input … and then reconstruct [the input] into its original sequential form” (Armelin [0019]) to ensure the data passed into the CNN-LSTM architecture is suitable input.
Mitchko/Lopes/Armelin fails to explicitly disclose but Kim discloses generating filtered data based on inputting the denoised features from the DAE to a convolutional neural network (CNN). (Kim [0059]; “The preprocessing unit 30 may include a raw image feature derivation module 31, a viewpoint image feature derivation module 32, and an input data generation module 33. The raw image feature derivation module 31 derives a 1D feature of the raw image for each frame by using a convolutional neural network (CNN) algorithm. That is, the CNN is an algorithm that generates feature maps by applying a filter to the input raw image.” where the CNN generating feature maps through filtering reads on providing filtered data)
generating extracted features based on inputting the filtered data from the CNN to at least one feature extraction layer (Kim [0060]; “and progressively reduces the sizes, thereby extracting only features”, wherein progressively reducing feature maps to extract features reads on multiple feature extraction layers providing extracted features)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin’s method of predicting net settlement positions in interactions based on received historical transaction data to use Kim’s model architecture utilizing CNN layers for filtering and feature extraction. The motivation to do so is to allow Mitchko/Lopes/Armelin’s model to generate cleaner input data for LSTM prediction when “deriving the user viewpoint image with no distortion” (Kim [0065]).
The combination of Mitchko/Lopes/Armelin/Kim teaches generating the at least one prediction of the net settlement position of the first entity. (Mitchko [0083]; “In one embodiment, the prediction comprises a prediction of a decrease or an increase in the future processing volume of the merchant”, wherein the predicted decreased or increased processing volume reads on a predicted net settlement position of an entity). The combination of Mitchko/Lopes/Armelin/Kim does not already teach but Kim further teaches generating the at least one prediction of the net settlement position of the first entity based on inputting at least one of the plurality of features or the extracted features to a long short-term memory (LSTM) model (Kim [0020]; “Preferably, the LSTM execution unit is a time-series deep-learning artificial neural network performing estimation of the viewpoint with respect to the input data. Preferably, the LSTM execution unit includes: an encoder extracting feature data for executing LSTM on the input data of the preprocessing unit, and transmitting the extracted feature data and a computed state vector; and a decoder estimating the viewpoint in the predetermined future by performing learning on the feature data and the computed state vector.”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin/Kim’s method of predicting net settlement positions in interactions based on received historical transaction data to additionally use Kim’s model architecture utilizing a LSTM for prediction. The motivation to do so is to allow Mitchko/Lopes/Armelin/Kim’s predictions to “receive a long-term state vector c.sub.t from the encoder 41 at every time t and performs learning so as to estimate the user viewpoint in the future t′ (Kim [0074]) thus retaining long-term dependencies across time.
Regarding Claim 4,
The combination of Mitchko/Lopes/Armelin/Kim teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Mitchko/Lopes/Armelin/Kim already discloses wherein the payment transaction time series data comprises at least one of historical transaction data, historical settlement position data, daily settlement data, real-time authorization data, or any combination thereof. (Mitchko [0033]; “The one or more inputs may comprise transaction data and/or historical datasets associated with a particular user. The historical datasets may comprise a pattern of the user's purchases and/or account balances over a particular period of time”)
Regarding Claim 10,
The combination of Mitchko/Lopes/Armelin/Kim teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). The combination already discloses wherein the DAE comprises a recurrent neural network (RNN) autoencoder. (Armelin [0021]; “In another example of the denoising apparatus, the neural network comprises a Long Short-Term Memory, LSTM, recurrent neural network or a Gated Recurrent Unit, GRU, recurrent neural network.”)
Regarding Claim 11,
The combination of Mitchko/Lopes/Armelin/Kim teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). The combination already discloses wherein the at least one feature extraction layer comprises at least one fully connected neural network layer. (Kim [0060]; The raw image feature derivation module 31 derives a 1D feature of the raw image for each frame by using a convolutional neural network (CNN) algorithm)
Regarding Claim 12,
The combination of Mitchko/Lopes/Armelin/Kim teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). The combination already discloses wherein inputting the at least one of the plurality of features or the extracted features into the LSTM model to provide the at least one prediction comprises inputting an output of the LSTM model to a sequence decoder to provide the at least one prediction. (Kim [0022]; “Preferably, the LSTM execution unit includes …
a decoder estimating the viewpoint in the predetermined future by performing learning on the feature data and the computed state vector.” wherein the estimated viewpoint reads on a prediction and a decoder for LSTMs, which are inherently sequence-based due to processing specifically sequential data, reads on a sequence decoder)
Claim 16 recites a system comprising a processor, as recited in the method of Claim 1, as well as a non-transitory computer-readable medium including instructions to execute the same method of Claim 1. As performance of an abstract idea on generic computing components (processor, non-transitory computer-readable medium including instructions) cannot integrate an abstract into a practical application nor provide significantly more than the abstract idea itself (see MPEP 2106.05(f)), Claim 16 is rejected for reasons set forth in the rejection of Claim 1.
Claim 20 recites a computer program product comprising a processor, as recited in the method of Claim 1, as well as a non-transitory computer-readable medium including instructions to execute the same method of Claim 1. As performance of an abstract idea on generic computing components (processor, non-transitory computer-readable medium including instructions) cannot integrate an abstract into a practical application nor provide significantly more than the abstract idea itself (see MPEP 2106.05(f)), and is thus rejected for reasons set forth in the rejection of Claim 1.
Claims 2-3 are rejected under 35 U.S.C. 103 as being anticipated by Mitchko et al. (US20220215465A1, hereinafter “Mitchko”) in view of Lopes et al. (US20210350382A1, hereinafter “Lopes”) in view of Armelin et al (US20220397425, hereinafter “Armelin”) in view of Kim et al. (US20220180469A1, hereinafter “Kim”) and further in view of Xu et al. (US20190244236A1, hereinafter “Xu”)
Regarding Claim 2,
The Mitchko/Lopes/Armelin/Kim combination teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Mitchko/Lopes/Armelin/Kim fails to explicitly disclose but Xu discloses wherein the payment transaction time series data associated with each transaction of the first subset of payment transactions comprises an account identifier associated with the first entity. (Xu [0022]; “an acquirer business identification number (BIN) may be selected from the bulk electronic transaction data 103. The BIN may be part of the transaction data 103 and may be selected to assist in limiting the more detailed review to merchants 107 that are part of the offer. The BIN may follow a known protocol and only a portion of the BIN may need to be reviewed to determine if further analysis is required”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin/Kim’s method of predicting net settlement positions in interactions based on received historical transaction data to associate each transaction datum with Xu’s account identifier. The motivation to do so lies in how “BIN may be used to pinpoint whether a merchant is a registered merchant 107” (Xu [0023]), thus allowing Mitchko/Lopes/Armelin/Kim to determine whether the entity of a given interaction is registered or unregistered in a given business database.
Regarding Claim 3,
The combination of Mitchko/Lopes/Armelin/Kim/Xu teaches the method of Claim 2 (and thus the rejection of Claim 2 is incorporated). The combination already discloses wherein the account identifier comprises at least one of a settlement reporting entity (SRE) number, a funds transfer SRE (FTSRE) number, a business identification (BID) number, or any combination thereof. (Xu [0022]; “an acquirer business identification number (BIN) may be selected from the bulk electronic transaction data 103. The BIN may be part of the transaction data 103 and may be selected to assist in limiting the more detailed review to merchants 107 that are part of the offer. The BIN may follow a known protocol and only a portion of the BIN may need to be reviewed to determine if further analysis is required”, wherein the BIN reads on a BID)
Claims 5-6 and 17 are rejected under 35 U.S.C. 103 as being anticipated by Mitchko et al. (US20220215465A1, hereinafter “Mitchko”) in view of Lopes et al. (US20210350382A1, hereinafter “Lopes”) in view of Armelin et al (US20220397425, hereinafter “Armelin”) in view of Kim et al. (US20220180469A1, hereinafter “Kim”) and further in view of Blanchard et al. (US20220067541, hereinafter “Blanchard”)
Regarding Claim 5,
The Mitchko/Lopes/Armelin/Kim combination teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Mitchko/Lopes/Armelin/Kim fails to explicitly disclose but Blanchard discloses wherein determining the plurality of features comprises determining the plurality of features based on a random forest model. (Blanchard [0035]; “The feature engineering module 220 ranks the features and determines a ranking score for each feature. The ranking score of a feature indicates how important the feature is for predicting the target variable, in other words, how good the feature is as a predictor. In some embodiments, the feature engineering module 220 constructs a random forest based on the features and the dataset. The feature engineering module 220 determines a ranking score of a feature based on each decision tree in the random forest and obtains an average of the individual ranking scores as the ranking score of the feature. The feature engineering module 220 may use GINI impurity as part of each decision tree to measure how much a feature contributes to the whole predictive model. The ranking score of a feature determined by using the random forest indicates how important the feature is relative to the other features and are referred to as “relative ranking score.” In one example, the ranking module 330 determines that the relative ranking scores of the highest ranked selected feature is 1. The ranking module 330 then determines a ratio of the ranking score of each of the rest of the features to the ranking score of the highest ranked feature as the relative ranking scores of the corresponding selected feature”
Blanchard [0036]; “The feature engineering module 220 may select a subset of the group of features based on their relative ranking scores and/or absolute ranking scores as features to train the model.” wherein the subset of features input to the model is determined through a random forest)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin/Kim’s method of predicting net settlement positions in interactions based on received historical transaction data to use Blanchard’s random forest model to determine features for input in Mitchko/Lopes/Armelin/Kim’s prediction model. The motivation to do so lies in enhancing Mitchko/Lopes/Armelin/Kim’s feature selection to consider “how important the feature is relative to the other features” (Blanchard [0035]) and thus only select the most important features.
Regarding Claim 6,
The combination of Mitchko/Lopes/Armelin/Kim/Blanchard teaches the method of Claim 5 (and thus the rejection of Claim 5 is incorporated). The combination already discloses wherein determining the plurality of features based on the random forest model comprises: receiving a first plurality of features; (Blanchard [0034]; “The feature engineering module 220 extracts features from the dataset. The feature engineering module 220 may extract a variable in the dataset as a feature and/or use a transformer to convert a variable in the dataset to a feature” where extracting features from the dataset reads on receiving a first plurality of features)
evaluating the first plurality of features with the random forest model to rank the first plurality of features based on a respective level of impact of each respective features of the first plurality of features on an output of the at least one machine learning model; (Blanchard [0035]; “The feature engineering module 220 ranks the features and determines a ranking score for each feature. The ranking score of a feature indicates how important the feature is for predicting the target variable, in other words, how good the feature is as a predictor. In some embodiments, the feature engineering module 220 constructs a random forest based on the features and the dataset. The feature engineering module 220 determines a ranking score of a feature based on each decision tree in the random forest and obtains an average of the individual ranking scores as the ranking score of the feature. The feature engineering module 220 may use GINI impurity as part of each decision tree to measure how much a feature contributes to the whole predictive model. The ranking score of a feature determined by using the random forest indicates how important the feature is relative to the other features and are referred to as “relative ranking score.” In one example, the ranking module 330 determines that the relative ranking scores of the highest ranked selected feature is 1. The ranking module 330 then determines a ratio of the ranking score of each of the rest of the features to the ranking score of the highest ranked feature as the relative ranking scores of the corresponding selected feature.”)
and selecting a second plurality of features based on ranking of the first plurality of features, wherein the second plurality of features comprises the plurality of features; (Blanchard [0036]; “The feature engineering module 220 may select a subset of the group of features based on their relative ranking scores and/or absolute ranking scores as features to train the model.”)
Regarding Claim 17,
Mitchko/Lopes/Armelin/Kim teaches the method of Claim 16 (and thus the rejection of Claim 16 is incorporated). Mitchko/Lopes/Armelin/Kim fails to explicitly disclose but Blanchard discloses wherein determining the plurality of features comprises determining the plurality of features based on a random forest model. (Blanchard [0035]; “The feature engineering module 220 ranks the features and determines a ranking score for each feature. The ranking score of a feature indicates how important the feature is for predicting the target variable, in other words, how good the feature is as a predictor. In some embodiments, the feature engineering module 220 constructs a random forest based on the features and the dataset. The feature engineering module 220 determines a ranking score of a feature based on each decision tree in the random forest and obtains an average of the individual ranking scores as the ranking score of the feature. The feature engineering module 220 may use GINI impurity as part of each decision tree to measure how much a feature contributes to the whole predictive model. The ranking score of a feature determined by using the random forest indicates how important the feature is relative to the other features and are referred to as “relative ranking score.” In one example, the ranking module 330 determines that the relative ranking scores of the highest ranked selected feature is 1. The ranking module 330 then determines a ratio of the ranking score of each of the rest of the features to the ranking score of the highest ranked feature as the relative ranking scores of the corresponding selected feature”
Blanchard [0036]; “The feature engineering module 220 may select a subset of the group of features based on their relative ranking scores and/or absolute ranking scores as features to train the model.” wherein the subset of features input to the model is determined through a random forest)
wherein determining the plurality of features based on the random forest model comprises: receiving a first plurality of features; (Blanchard [0034]; “The feature engineering module 220 extracts features from the dataset. The feature engineering module 220 may extract a variable in the dataset as a feature and/or use a transformer to convert a variable in the dataset to a feature” where extracting features from the dataset reads on receiving a first plurality of features)
evaluating the first plurality of features with the random forest model to rank the first plurality of features based on a respective level of impact of each respective features of the first plurality of features on an output of the at least one machine learning model; (Blanchard [0035]; “The feature engineering module 220 ranks the features and determines a ranking score for each feature. The ranking score of a feature indicates how important the feature is for predicting the target variable, in other words, how good the feature is as a predictor. In some embodiments, the feature engineering module 220 constructs a random forest based on the features and the dataset. The feature engineering module 220 determines a ranking score of a feature based on each decision tree in the random forest and obtains an average of the individual ranking scores as the ranking score of the feature. The feature engineering module 220 may use GINI impurity as part of each decision tree to measure how much a feature contributes to the whole predictive model. The ranking score of a feature determined by using the random forest indicates how important the feature is relative to the other features and are referred to as “relative ranking score.” In one example, the ranking module 330 determines that the relative ranking scores of the highest ranked selected feature is 1. The ranking module 330 then determines a ratio of the ranking score of each of the rest of the features to the ranking score of the highest ranked feature as the relative ranking scores of the corresponding selected feature.”)
and selecting a second plurality of features based on ranking of the first plurality of features, wherein the second plurality of features comprises the plurality of features; (Blanchard [0036]; “The feature engineering module 220 may select a subset of the group of features based on their relative ranking scores and/or absolute ranking scores as features to train the model.”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin/Kim’s method of predicting net settlement positions in interactions based on received historical transaction data to use Blanchard’s random forest model to determine features for input in Mitchko/Lopes/Armelin/Kim’s prediction model. The motivation to do so lies in enhancing Mitchko/Lopes/Armelin/Kim’s feature selection to consider “how important the feature is relative to the other features” (Blanchard [0035]) and thus only select the most important features.
Claims 7-8 and 18 are rejected under 35 U.S.C. 103 as being anticipated by Mitchko et al. (US20220215465A1, hereinafter “Mitchko”) in view of Lopes et al. (US20210350382A1, hereinafter “Lopes”) in view of Armelin et al (US20220397425, hereinafter “Armelin”) in view of Kim et al. (US20220180469A1, hereinafter “Kim”) and further in view of Han et al. (US20230028574A1, hereinafter “Han”)
Regarding Claim 7,
The combination of Mitchko/Lopes/Armelin/Kim teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Mitchko/Lopes/Armelin/Kim fails to explicitly disclose but Han discloses wherein at least one of the plurality of machine learning models comprises at least one of an additive regression model, a Prophet model, or any combination thereof. (Han [0056]; “Herein, the time series prediction model may be one of a prophet model, an autoregressive model, a moving average model or an autoregressive moving average model. In an exemplary embodiment, the prophet model is used to fit each of the component series.”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin/Kim’s method of predicting net settlement positions in interactions based on received historical transaction data to use Han’s Prophet models over the combination’s CNN models for predictions. The motivation to do so lies in how “when adopting the prophet model … prediction may be carried out according to the trend of periodic changes and aperiodic changes of the traffic data and the impact of holidays on the traffic data, which improves accuracy” (Han [0101]).
Regarding Claim 8,
The combination of Mitchko/Lopes/Armelin/Kim/Han teaches the method of Claim 7 (and thus the rejection of Claim 7 is incorporated). The combination already discloses wherein the Prophet model comprises the additive regression model comprising at least one of: a piecewise linear or logistic growth curve trend; a yearly seasonal component modeled using Fourier series; a weekly seasonal component; a list of holidays; or any combination thereof. (Han [0101]; “The prophet model automatically detects changes in trends by selecting changepoints from the data. Further, a yearly seasonal component is modeled using the Fourier series. Furthermore, a weekly seasonal component modeled using dummy variables”)
Regarding Claim 18,
The combination of Mitchko/Lopes/Armelin/Kim teaches the method of Claim 16 (and thus the rejection of Claim 16 is incorporated). Mitchko/Lopes/Armelin/Kim fails to explicitly disclose but Han discloses wherein at least one of the plurality of machine learning models comprises at least one of an additive regression model, a Prophet model, or any combination thereof; (Han [0056]; “Herein, the time series prediction model may be one of a prophet model, an autoregressive model, a moving average model or an autoregressive moving average model. In an exemplary embodiment, the prophet model is used to fit each of the component series.”)
And wherein the Prophet model comprises the additive regression model comprising at least one of: a piecewise linear or logistic growth curve trend; a yearly seasonal component modeled using Fourier series; a weekly seasonal component; a list of holidays; or any combination thereof. (Han [0101]; “The prophet model automatically detects changes in trends by selecting changepoints from the data. Further, a yearly seasonal component is modeled using the Fourier series. Furthermore, a weekly seasonal component modeled using dummy variables”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin/Kim’s method of predicting net settlement positions in interactions based on received historical transaction data to use Han’s Prophet models over Mitchko/Lopes/Armelin/Kim’s base CNN models for predictions. The motivation to do so lies in how “when adopting the prophet model … prediction may be carried out according to the trend of periodic changes and aperiodic changes of the traffic data and the impact of holidays on the traffic data, which improves accuracy” (Han [0101]).
Claims 13-14 are rejected under 35 U.S.C. 103 as being anticipated by Mitchko et al. (US20220215465A1, hereinafter “Mitchko”) in view of Lopes et al. (US20210350382A1, hereinafter “Lopes”) in view of Armelin et al (US20220397425, hereinafter “Armelin”) in view of Kim et al. (US20220180469A1, hereinafter “Kim”) and further in view of McDonald et al. (US11645617B1, hereinafter “McDonald”)
Regarding Claim 13,
The combination of Mitchko/Lopes/Armelin/Kim teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Mitchko/Lopes/Armelin/Kim fails to explicitly disclose but McDonald discloses wherein the at least one prediction comprises a plurality of predictions comprising a respective prediction of the net settlement position of the first entity for each subperiod of a time period. (McDonald [0047]; “Prediction module 270 applies samples of transaction data 280, supply chain data 282, product data 284, inventory data 286, store data 290, customer data 292, demand forecasts 294, and other data to prediction models 298 to generate predictions and calculated factor values for one or more causal factors. According to embodiments, prediction module 270 may predict a volume Y (target or label) from a set of causal factors X along with causal factors strengths that describe the strength of each causal factor variable contributing to the predicted volume. According to some embodiments, prediction module 270 generates predictions at daily intervals. However, embodiments contemplate longer and shorter prediction phases that may be performed, for example, weekly, twice a week, twice a day, hourly, or the like.”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin/Kim’s method of predicting net settlement positions in interactions based on received historical transaction data to use McDonald’s prediction generation at subperiods. The motivation to do so is to modify Mitchko/Lopes/Armelin/Kim’s predictions to accommodate for fluctuating “training data [comprising] historic sales patterns, prices, promotions, weather conditions, and other factors influencing future demand … on a specific day” (McDonald [0029])
Regarding Claim 14,
The combination of Mitchko/Lopes/Armelin/Kim/McDonald teaches the method of Claim 13 (and thus the rejection of Claim 13 is incorporated). The combination already discloses wherein the time period comprises seven days, each subperiod comprises one day of the seven days, and the plurality of predictions comprises a first prediction for a first day of the seven days, a second prediction for a second day of the seven days, a third prediction for a third day of the seven days, a fourth prediction for a fourth day of the seven days, a fifth prediction for a fifth day of the seven days, a sixth prediction for a sixth day of the seven days, and a seventh prediction for a seventh day of the seven days. (McDonald [0047]; “According to some embodiments, prediction module 270 generates predictions at daily intervals. However, embodiments contemplate longer and shorter prediction phases that may be performed, for example, weekly, twice a week, twice a day, hourly, or the like.”)
Claim 15 is rejected under 35 U.S.C. 103 as being anticipated by Mitchko et al. (US20220215465A1, hereinafter “Mitchko”) in view of Lopes et al. (US20210350382A1, hereinafter “Lopes”) in view of Armelin et al (US20220397425, hereinafter “Armelin”) in view of Kim et al. (US20220180469A1, hereinafter “Kim”) and further in view of Puthiyapurayil et al. (US11645611B1, hereinafter “Puthiyapurayil”)
Regarding Claim 15,
The combination of Mitchko/Lopes/Armelin/Kim teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Mitchko/Lopes/Armelin/Kim fails to explicitly disclose but Puthiyapurayil discloses wherein communicating the at least one prediction comprises communicating the at least one prediction to the first entity system via at least one of a graphical user interface (GUI) or an application programming interface (API). (Puthiyapurayil [0071]; “User interface module 250 of planning and execution system 130n generates and displays a UI, such as, for example, a GUI, that displays one or more interactive visualizations of transaction data 260, supply chain data 262, product data 264, inventory data 266, inventory policies 268, store data 270, customer data 272, supply chain models 274, archived trained machine-learning models 276, and prediction data”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify Mitchko/Lopes/Armelin/Kim’s method of predicting net settlement positions in interactions based on received historical transaction data to use Puthiyapurayil’s GUI to communicate prediction results. The motivation to do so is to “display one or more interactive visualizations of … prediction data” [Puthiyapurayil 0071] to visually aid user’s understanding of predictions.
Response to Arguments
The Examiner acknowledges the Applicant’s amendments in which Claims 1, 2, 4, 16 and 20 are amended.
Applicant’s arguments filed February 2nd, 2026, traversing the rejection of claims 1-8, 10-18 and 20 under 35 U.S.C. § 101 have been fully considered and are fully persuasive.
Applicant’s arguments regarding the 35 U.S.C. § 103 rejection of claims 1-8, 10-18 and 20 of the previous office action have been considered, but are not fully persuasive.
Applicant recites on pages 10-11 of remarks that the cited art, whether taken separately or in combination, does not disclose or suggest the method as claimed in the amended Claim 1.
Examiner respectfully disagrees. The combination of Mitchko/Lopes/Armelin/Kim already discloses the amended limitations of the plurality of payment transactions being in a time series. The historical datasets disclosed by Mitchko comprising transactional data, spending patterns, and general consumer behavior over a select period of time is interpretable as the payment transaction data being of a time series nature.
Mitchko discloses receiving payment transaction time series data and generation of a net settlement position. Mitchko was combined with Lopes to disclose feature extraction performed on Mitchko’s received payment transaction time series data, thus disclosing generation of a net settlement position based on Lopes’ extracted features of Mitchko’s received data instead of Mitchko’s received data directly. Armelin then discloses a denoising autoencoder taking as input a plurality of features. By inputting Mitchko/Lopes’ extracted plurality of features into Armelin’s denoising autoencoder, the combination thus discloses generating denoised feature based on inputting the plurality of features to a denoising autoencoder. Kim discloses a convolutional neural network taking as input denoised features, as well as a feature extraction layer generating finally extracted features taking as input the filtered CNN data. By inputting Mitchko/Lopes/Armelin’s DAE denoised extracted features into Kim’s CNN, the combination thus discloses generating filtered data based on inputting the denoised features from the DAE to a convolutional neural network. By taking such generated filtered data and further inputting it into Kim’s disclosed feature extraction layer, the combination also discloses generating extracted features based on inputting the filtered data from the CNN to at least one feature extraction layer. Kim also discloses a long short-term memory model which receives a plurality of features or extracted features to generate a prediction. By inputting the combination’s extracted features into Kim’s LSTM, the combination thus discloses generating the at least one prediction of the net settlement position of the first entity based on inputting at least one of the plurality of features or the extracted features to a long short-term memory (LSTM) model. Thus, the combination of Mitchko/Lopes/Armelin/Kim demonstrably discloses the amended method as claimed in its entirety.
The rejection of Claim 1 under 35 U.S.C. § 103 has been maintained. Similarly, the rejection of Claims 16 and 20 under 35 U.S.C. § 103 have been maintained.
The rejection of Claims 2-8 and 10-15 under 35 U.S.C. § 103, which depend directly or indirectly from Claim 1, have been maintained.
The rejection of Claims 17 and 18 under 35 U.S.C. § 103, which depend directly or indirectly from Claim 16, have been maintained.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JONATHAN J KIM whose telephone number is (571) 272-0523. The examiner can normally be reached 9-6.
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, Matt El can be reached on (571) 270-3264. 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
/JONATHAN J KIM/Examiner, Art Unit 2141
/MATTHEW ELL/Supervisory Patent Examiner, Art Unit 2141
Prosecution Insights