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 .
Status of Claims
This action is in reply to the communications filed on 06/12/2025.
Claims 1-12 are currently pending and have been examined.
This action is made Non-Final.
Priority
Application 19/236,908 (instant application) was filed on June 12th, 2025, which is a continuation of parent Application 18/337,350 (abandoned) filed on June 19th, 2023.
Related Application(s) - Prior Art of Record
The instant application is a continuation application (CON) of parent Application 18/337,350. In accordance with MPEP§609.02 A.2 and §2001.06(b) (last paragraph), the prior art cited in the above parent application has been considered, and all documents cited or considered 'of record' in that application are now considered cited or 'of record' in this application. The prosecution history of the above parent application is relevant in the examination of the instant application.
Examiner’s Note
If abandoned application, 18/337,350 is revived, an obviousness double patenting rejection would be applicable.
Examiner Request
The Applicant is requested to indicate where in the specification there is support for future claim amendments to avoid U.S.C 112(a) issues that can arise. The Examiner thanks the Applicant in advance.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 06/12/2025 was filed before the mailing date of a first Office Action on the merits. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Objections
Claims 1 and 7 are objected to because of the following informalities:
Claim 1: line 16 and Claim 7: line 8 recites the limitation “data fields.” Claim 1: line 12 and Claim 7: line 6 recite “financial data fields.” “Data fields” and “financial data fields” are interchangeably used. Is “data fields” recited in Claim 1: line 16 and Claim 7: line 8 different that the “financial data fields” recited in Claim 1: line 12 and Claim 7: line 6? It appears there is a typographical mistake since the specification only points to a single instance of data fields for this interpretation. For compact examination purposes, Examiner interpreted the instances recited in Claim 1: line 16 and Claim 7: line 8 as “financial data fields.” Appropriate correction is required.
Furthermore, Claim 1: line 16 and Claim 7: line 8 recites the limitation “extracting data fields .” As discussed above, “data fields” is initially and previously recited in Claim 1: line 12 and Claim 7: line 6. Is the ‘data fields’ recited in Claim 1: line 16 and Claim 7: line 8 different than ‘data fields’ initially and previously recited in Claim 1: line 12 and Claim 7: line 6? It appears there is a typographical mistake since the specification only points to a single instance of data fields for this interpretation. For compact examination purposes, Examiner interpreted the instances recited in Claim 1: line 16 and Claim 7: line 8 as “extracting the [financial] data fields.” Appropriate correction is required.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-12 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea of receiving and processing financial data associated with a borrower and generating a confidence score in response to the query; without significantly more.
Examiner has identified claim 1 as the claim that represents the claimed invention presented in independent claims 1 and 7.
Claim 1 is directed to a system which is one of the statutory categories of invention; (Step 1: YES); and Claim 7 is directed to a method which is one of the statutory categories of invention (Step 1: YES).
Claim 1 is directed to a system comprising: a computing device comprising a memory and a processor; a data acquisition engine comprising a first plurality of programming instructions stored in the memory which, when operating on the processor, causes the computing device to: receive one or more documents associated with a borrower; feed the one or more documents into a first machine learning model comprising a convolutional neural network configured to: normalize documents of varying dimensions using adaptive pooling; extract multi-scale features through a plurality of convolutional layers; apply a spatial attention mechanism to identify and weight document regions containing financial data fields; and output document classification and associated confidence scores; feed each of the one or more documents and its classification into a second machine learning model configured to validate the data by: extracting data fields using classification-specific parsing patterns; detecting anomalous values using a trained autoencoder that compares reconstruction error against learned thresholds; performing cross-document verification by mapping relationships between related financial fields; and generating field-level validation confidence scores; store the validated data and confidence scores in a borrower profile; and a generative artificial intelligence model configured to: receive as input a query and the borrower profile including the validation confidence scores; and generate predictive responses to the query weighted by the validation confidence scores. These limitations describe the abstract idea of receiving and processing financial data associated with a borrower and generating a confidence score in response to the query (with exception of the italicized and bolded terms above), which is mitigating risk of potential bias in loans offered by lenders to borrowers, and verifying and validating a borrower’s data and confidence scores for a lending institution; therefore, corresponding to a fundamental economic principle or practice (including mitigating risk). Hence, a fundamental economic principle or practice (mitigating risk) is a Certain Methods of Organizing Human Activity. The abstract idea is also the processing of a borrower’s data and confidence scores to originate a loan between a borrower and lending financial institution, which is commercial interaction. Therefore, a commercial interaction is also a Certain Methods of Organizing Human Activity. The computer device limitations, e.g., a computing device, memory, processor, data acquisition engine, first plurality of programming instructions, first machine learning model, convolutional neural network, plurality of convolutional layers, spatial attention mechanism, second machine learning model, trained autoencoder, and generative artificial intelligence model, do not necessarily restrict the claim from reciting an abstract idea. Thus, claim 1 recites an abstract idea (Step 2A-Prong 1: YES).
This judicial exception is not integrated into a practical application because the additional elements of a computing device, memory, processor, data acquisition engine, first plurality of programming instructions, first machine learning model, convolutional neural network, plurality of convolutional layers, spatial attention mechanism, second machine learning model, trained autoencoder, and generative artificial intelligence model, are no more than simply applying the abstract idea using generic computer elements. The additional elements listed above are all recited at a high level of generality and under their broadest reasonable interpretation comprises a generic computing arrangement. Merely invoking a computing device, memory, processor, data acquisition engine, first plurality of programming instructions, first machine learning model, convolutional neural network, plurality of convolutional layers, spatial attention mechanism, second machine learning model, trained autoencoder, and generative artificial intelligence model is similar to invoking software and software components. The presence of a generic computer arrangement is nothing more than to implement the claimed invention (MPEP 2106.05(f)). The computer network limitations are a field of use limitations (MPEP 2106.05(h)). Therefore, the recitations of additional elements do not meaningfully apply the abstract idea and hence do not integrate the abstract idea into a practical application. Thus, claim 1 does not integrate the abstract idea into a practical application (Step 2A-Prong 2: NO).
Claim 1 does not include additional elements that are sufficient to amount to significantly more than the judicial exception because the additional elements of a computing device, memory, processor, data acquisition engine, first plurality of programming instructions, first machine learning model, convolutional neural network, plurality of convolutional layers, spatial attention mechanism, second machine learning model, trained autoencoder, and generative artificial intelligence model are recited at a high level of generality in that it results in no more than simply applying the abstract idea using generic computer elements. The computer network limitations are a field of use limitations (MPEP 2106.05(h)). The additional elements when considered separately and as an ordered combination do not amount to add significantly more as these limitations provide nothing more than to simply apply the exception in a generic computer environment (Step 2B: NO). Thus, claim 1 is not patent eligible.
Similar arguments can be extended to the other independent claim, claim 7; and hence claim 7 is rejected on similar grounds as claim 1.
Dependent Claims 2-6 and 8-12 are directed to a system and method, respectively, which perform steps that describe the abstract idea of receiving and processing financial data associated with a borrower and generating a confidence score in response to the query. Specifically, dependent claims 2, 3, 6, 8, 9, and 12 are directed to a system and a method, respectively, which perform the following steps: “wherein the first machine learning model is a trained classifier network; wherein the second machine learning model is trained using a regression algorithm; further comprising an application programming interface comprising a second plurality of programming instructions stored in the memory which, when operating on the processor, causes the computing device to: transmit the validated data in the borrower profile to a loan origination system associated with the one or more authorized lender institutions; and wherein the plurality of convolutional layers includes three layers for three different granularities,” (with exception of the italicized and bolded terms above), which is mitigating risk of potential bias in loans offered by lenders to borrowers, and verifying and validating a borrower’s data and confidence scores for a lending institution; therefore, corresponding to a fundamental economic principle or practice (including mitigating risk). Hence, a fundamental economic principle or practice (mitigating risk) is a Certain Methods of Organizing Human Activity. The abstract idea is also the processing of a borrower’s data and confidence scores to originate a loan between a borrower and lending financial institution, which is commercial interaction. Therefore, a commercial interaction is also a Certain Methods of Organizing Human Activity. The system limitations, e.g., a computing device, memory, processor, data acquisition engine, first plurality of programming instructions, first machine learning model, convolutional neural network, plurality of convolutional layers, spatial attention mechanism, second machine learning model, trained autoencoder, generative artificial intelligence model, trained classifier network, regression algorithm, application programming interface, second plurality of programming instructions, loan origination system, and three layers, do not necessarily restrict the claim from reciting an abstract idea. Thus, claims 2-6 and 8-12 are directed to an abstract idea. The additional limitations of a computing device, memory, processor, data acquisition engine, first plurality of programming instructions, first machine learning model, convolutional neural network, plurality of convolutional layers, spatial attention mechanism, second machine learning model, trained autoencoder, generative artificial intelligence model, trained classifier network, regression algorithm, application programming interface, second plurality of programming instructions, loan origination system, and three layers are no more than simply applying the abstract idea using generic computer elements. The additional elements listed above are all recited at a high level of generality and under their broadest reasonable interpretation comprises a generic computing arrangement. Merely invoking a computing device, memory, processor, data acquisition engine, first plurality of programming instructions, first machine learning model, convolutional neural network, plurality of convolutional layers, spatial attention mechanism, second machine learning model, trained autoencoder, generative artificial intelligence model, trained classifier network, regression algorithm, application programming interface, second plurality of programming instructions, loan origination system, and three layers is similar to invoking software and software components. The presence of a generic computer arrangement is nothing more than to implement the claimed invention (MPEP 2106.05(f)). The computer network limitations are a field of use limitations (MPEP 2106.05(h)). Therefore, the recitations of additional elements do not meaningfully apply the abstract idea, and hence do not integrate the abstract idea into a practical application. Furthermore, the additional elements of a computing device, memory, processor, data acquisition engine, first plurality of programming instructions, first machine learning model, convolutional neural network, plurality of convolutional layers, spatial attention mechanism, second machine learning model, trained autoencoder, generative artificial intelligence model, trained classifier network, regression algorithm, application programming interface, second plurality of programming instructions, loan origination system, and three layers do not amount to add significantly more as these limitations provide nothing more than to simply apply the exception in a generic computer environment.
Dependent claims 2-6 and 8-12 have further defined the abstract idea that is present in their respective independent claims 1 and 7; and thus, correspond to Certain Methods of Organizing Human Activity, and hence are abstract in nature for the reason presented above. The computer network limitations are a field of use limitations (MPEP 2106.05(h)). The dependent claims 2-6 and 8-12 do not include any additional elements that integrate the abstract idea into a practical application or are sufficient to amount to significantly more than the judicial exception when considered both individually and as an ordered combination. Therefore, claims 1-12 are directed to an abstract idea without significantly more.
Thus, claims 1-12 are not patent-eligible.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 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 set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied 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 1-12 are rejected under 35 U.S.C. 103 as being unpatentable over James (U.S. Patent Application Publication No. US 2023/0009149 A1 hereinafter “James”), in view of Paramasamy (U.S. Patent Publication No. US 11,561,666 B1; hereinafter “Paramasamy”).
Regarding Claims 1 and 7:
James teaches:
A system for intelligent document processing with anomaly detection and predictive analysis, (James, A system and method for processing loans includes a machine learning model associated with processing loans. The machine learning model may be configured based on an objective function associated with loan processing. One or more weights of the objective function may be updated to account for changes in one or more business conditions. The machine learning model may be configured based on the updates to the one or more weights. (See, Abstract); the proving or disproving of the validity of the model, and the employment of the predictive analytics to run scenarios that will help guide future actions. The goal of statistical analysis is to identify features that are unique (See, 34-38));
comprising: a computing device comprising a memory and a processor; (James, a module may be implemented as a hardware circuit or in programmable hardware devices. Modules may also be implemented in software for execution by various types of processors. Modules or portions of a module that are implemented in software, may be stored on one or more computer readable storage media (See, Para. 27, 88, 91));
a data acquisition engine comprising a first plurality of programming instructions stored in the memory which, when operating on the processor, causes the computing device to: (James, The environment 100 may include a loan approval and processing system 101 having an underwriting module 103 (See, Para. 28; Fig. 1: element 103));
receive one or more documents associated with a borrower; (James, The loan approval and processing system 101 will also receive input from a customer application subsystem 157 (See, Para. 28));
feed the one or more documents into a first machine learning model comprising a convolutional neural network configured to: (James, the relevance of the weighted features (Objective function 301) would be tested by the machine learning model development module 115 using machine learning models (e.g. ML model A 303, ML model B 305 and ML model n 307 implemented by the machine learning module 105. the results of such tests are used to train the machine learning module 105. For example, when input is received from machine learning processing subsystem 107 (Preprocessing and feature engineering), ML model A 303 may use a gradient boosting machine algorithm…. regression algorithms such as …. neural networks may be used for supervised learning. (See, Para. 47, 49));
normalize documents of varying dimensions using adaptive pooling; (James, Machine learning module 105 pre-processes the data in pre-processing subsystem 109. Pre-processing subsystem 109 pre-processes the data by formatting, cleaning, and sampling the data. The formatting step converts the data into a format that is suitable for use by the machine learning module 105….Finally standardizing numeric features (scaling values to lie within the same range) is necessary in regularized logistic regression algorithms, and the automated feature engineering subsystem 111 adds the steps because it knows that this step constitutes best practice. (See, Para. 32-35; Fig. 1));
extract multi-scale features through a plurality of convolutional layers; (James, If the calculated features do not clearly expose the predictive signals the utility of the models will be compromised. Feature engineering is the process for extracting numeric features. Thus, feature engineering is the process of transforming raw data into features that better represent the underlying problem to the predictive models, resulting in improved model accuracy on unseen data (See, Para. 32-35; Fig. 1); the relevance of the weighted features (Objective function 301) would be tested by the machine learning model development module 115 using machine learning models (e.g. ML model A 303, ML model B 305 and ML model n 307 implemented by the machine learning module 105. the results of such tests are used to train the machine learning module 105. For example, when input is received from machine learning processing subsystem 107 (Preprocessing and feature engineering), ML model A 303 may use a gradient boosting machine algorithm…. regression algorithms such as …. neural networks may be used for supervised learning. (See, Para. 47, 49));
apply a spatial attention mechanism to identify and weight document regions containing financial data fields; and (James, Machine learning module 105 pre-processes the data in pre-processing subsystem 109. Pre-processing subsystem 109 pre-processes the data by formatting, cleaning, and sampling the data. The formatting step converts the data into a format that is suitable for use by the machine learning module 105…... Similarly regularized logistic regression algorithms do not work with missing numerical values. In the credit risk model for example with a field containing the number of years since the applicant was last declared bankrupt, the field will be empty for most applicants. The automated feature engineering subsystem 111 would know that it should carry out missing values imputation as a feature engineering step. (See, Para. 32-35; Fig. 1, 5));
feed each of the one or more documents and its classification into a second machine learning model configured to validate the data by: extracting data fields using classification-specific parsing patterns; (James, the relevance of the weighted features (Objective function 301) would be tested by the machine learning model development module 115 using machine learning models (e.g. ML model A 303, ML model B 305 and ML model n 307 implemented by the machine learning module 105. the results of such tests are used to train the machine learning module 105. For example, when input is received from machine learning processing subsystem 107 (Preprocessing and feature engineering), ML model A 303 may use a gradient boosting machine algorithm….. logistic regression is a classification algorithm used to assign observations to a discrete set of classes. Unlike linear regression which outputs continuous number values, logistic regression transforms its output using the logistic sigmoid function to return a probability value which can then be mapped to two or more discrete classes. ML model n 307 may use other algorithms (See, Para. 32-35, 47-50));
extracting data fields using classification-specific parsing patterns; (James, Pre-processing subsystem 109 may determine simulated, estimated, or projected data to fill-in or complete data from a user based on the data from the user, by recognizing patterns in the data, fitting one or more functions to the data, or the like. The pre-processing subsystem 109 may fill in missing data using permutations of the missing data (e.g., each possible data value, each value at fixed increments between minimum and maximum values, or the like)……If the calculated features do not clearly expose the predictive signals the utility of the models will be compromised. Feature engineering is the process for extracting numeric features. Thus, feature engineering is the process of transforming raw data into features that better represent the underlying problem to the predictive models, resulting in improved model accuracy on unseen data (See, Para. 32-35; Fig. 1));
detecting anomalous values using a trained autoencoder that compares reconstruction error against learned thresholds; (James, Pre-processing subsystem 109 may determine simulated, estimated, or projected data to fill-in or complete data from a user based on the data from the user, by recognizing patterns in the data, fitting one or more functions to the data, or the like. The pre-processing subsystem 109 may fill in missing data using permutations of the missing data (e.g., each possible data value, each value at fixed increments between minimum and maximum values, or the like (See, Para. 32-35; Fig. 1, 5); the proving or disproving of the validity of the model, and the employment of the predictive analytics to run scenarios that will help guide future actions. The goal of statistical analysis is to identify features that are unique (See, 34-38));
performing cross-document verification by mapping relationships between related financial fields; and (James, Machine learning module 105 pre-processes the data in pre-processing subsystem 109. Pre-processing subsystem 109 pre-processes the data by formatting, cleaning, and sampling the data. The formatting step converts the data into a format that is suitable for use by the machine learning module 105…... Similarly regularized logistic regression algorithms do not work with missing numerical values. In the credit risk model for example with a field containing the number of years since the applicant was last declared bankrupt, the field will be empty for most applicants. The automated feature engineering subsystem 111 would know that it should carry out missing values imputation as a feature engineering step……model creation and testing subsystem 115 also includes model testing subsystem 119 that implements automated model testing to verify that derived models remain valid, and triggers relearning of a new model upon model failure. (See, Para. 32-36, 37-38, 50; Fig. 1, 5));
generating field-level validation confidence scores; (James, To assess model performance the data can be partitioned into a training set and a validation or test set. Training set used to construct the classifiers and the test set is used to assess their performance. The performance of each classifier on test set is used to obtain an unbiased estimator of the classifier's performance. If a number of classifiers are tested, then in the model selection stage the machine learning model development module 115 can choose the classifier that performed best on the test set. Performance may be determined by using a receiver operating characteristic curve that is created by plotting the true positive rate against the false positive rate at various threshold setting. (See, Para. 32-36, 37-38; Fig. 1, 5));
James does not specifically teach that to output document classification and associated confidence scores; store the validated data and confidence scores in a borrower profile; and a generative artificial intelligence model configured to: receive as input a query and the borrower profile including the validation confidence scores; and generate predictive responses to the query weighted by the validation confidence scores.
However, Paramasamy teaches the following limitations:
output document classification and associated confidence scores; (Paramasamy, the learning models can use Bayesian networks and/or other machine learning algorithms to identify new and statistically significant data associations and apply statistically calculated confidence scores to those associations, whereby confidence scores that do not meet a predetermined minimum threshold are eliminated. Bayesian networks are algorithms that can describe relationships or dependencies between certain variables. The algorithms calculate a conditional probability that an outcome is highly likely given specific evidence. As new evidence and outcome dispositions are fed into the algorithm, more accurate conditional probabilities are calculated that either prove or disprove a particular hypothesis. A Bayesian network essentially learns over time. In examples of the disclosed systems and methods, other machine learning algorithms can be selected, including, for example, linear regression, logistic regression, support vector machines, and neural networks. (See, Col.6: lines 11-28); The automated action can also be a prediction based on the proximity of the new loan segment to existing loan context models. For example, the automated action can be a prediction as to whether a loan will or not be underwritten in a future stage of the loan application process. (See, Col.16: lines 13-63; Col.17: lines 43-63; Col.18: lines 13-29); the system to: generate a contextual model for each of a plurality of context types, each of the contextual models defining a multi-dimensional virtual space; receive data associated with a loan application; parse the data into individual data elements…..virtually position the individual data elements in the multidimensional virtual space of the contextual model corresponding to the context type matching the individual elements with corresponding clusters of learned data elements for the associated context type; and based on the position, apply to the received data one or more rules associated with the corresponding clusters, causing the system to: automatically predict whether the loan application will be accepted or rejected for underwriting; and based on the automatically predict, automatically perform an action associated with the loan application. (See, Col. 1: lines 63-67; Col. 2: lines 1-17; Col.3: lines 26-53); At the modeling phase 306, segments of prepared data elements of the acquired data are created building clusters of elements in intelligent ways for loan application workplans using context modeling algorithms 216 and incorporating and positioning the newly acquired data elements relative to those clusters to create segments …. semantic similarity of data elements as described above can be used to create intelligent clustering of data elements (Col. 10: lines 25-47; Col. 13: lines 39-63)).
store the validated data and confidence scores in a borrower profile; and (Paramasamy, the interface 800 also includes a loan information pane 804 that provides a pictorial representation of the status of 25 the transaction. In this instance, the credit profile for the borrower is shown, including credit, income, assets, etc. The pane 804 includes example circle and line bars to illustrate the relative completeness of each depicted item associated with the customer's credit profile….as the information about the assets and other conditions change over time, the quantification can increase or decrease, with a 100 percent qualification profile indicating that the customer is approved for the loan transaction. It is appreciated that pane 804 is customized specifically to the customer, and the information shown and presented will dynamically change based on the particular customer. (Col.7: lines 38-67; Col. 9: lines 9-45; Col. 23: lines 24-43));
and a generative artificial intelligence model configured to: receive as input a query and the borrower profile including the validation confidence scores; and (Paramasamy, the data obtained at each stage is contextually classified according to an existing context model of an impacted workplan and/or used to create a new context model and/or to update an existing context model. Depending on the classification of data and the confidence in the classification using machine learning algorithms, the data is automatically incorporated into the loan application process and/or used to automatically trigger one or more actions. (See, Col. 7: lines 61-67; Col. 8: line 1); the step 408 can include associating the received document's data types ( e.g., the data of the segment representing the received document) with a particular contextual model or multiple contextual models…… Machine learning can be employed over time to automatically determine whether a received document has complete information with sufficiently high confidence, e.g., 100 percent confidence (Col. 15: lines 1-19));
generate predictive responses to the query weighted by the validation confidence scores. (Paramasamy, the learning models can use Bayesian networks and/or other machine learning algorithms to identify new and statistically significant data associations and apply statistically calculated confidence scores to those associations, whereby confidence scores that do not meet a predetermined minimum threshold are eliminated. Bayesian networks are algorithms that can describe relationships or dependencies between certain variables. The algorithms calculate a conditional probability that an outcome is highly likely given specific evidence. As new evidence and outcome dispositions are fed into the algorithm, more accurate conditional probabilities are calculated that either prove or disprove a particular hypothesis. A Bayesian network essentially learns over time. In examples of the disclosed systems and methods, other machine learning algorithms can be selected, including, for example, linear regression, logistic regression, support vector machines, and neural networks. (See, Col.6: lines 11-28); The automated action can also be a prediction based on the proximity of the new loan segment to existing loan context models. For example, the automated action can be a prediction as to whether a loan will or not be underwritten in a future stage of the loan application process. (See, Col.16: lines 13-63; Col.17: lines 43-63; Col.18: lines 13-29); the system to: generate a contextual model for each of a plurality of context types, each of the contextual models defining a multi-dimensional virtual space; receive data associated with a loan application; parse the data into individual data elements…..virtually position the individual data elements in the multidimensional virtual space of the contextual model corresponding to the context type matching the individual elements with corresponding clusters of learned data elements for the associated context type; and based on the position, apply to the received data one or more rules associated with the corresponding clusters, causing the system to: automatically predict whether the loan application will be accepted or rejected for underwriting; and based on the automatically predict, automatically perform an action associated with the loan application. (See, Col. 1: lines 63-67; Col. 2: lines 1-17; Col.3: lines 26-53); At the modeling phase 306, segments of prepared data elements of the acquired data are created building clusters of elements in intelligent ways for loan application workplans using context modeling algorithms 216 and incorporating and positioning the newly acquired data elements relative to those clusters to create segments …. semantic similarity of data elements as described above can be used to create intelligent clustering of data elements (Col. 10: lines 25-47; Col. 13: lines 39-63)).
It would have been obvious to one of ordinary skill in the art b before the effective filing of the claimed invention to have modified James with the features of Paramasamy’s system because “the present disclosure is directed to systems and methods that employ data driven decision automation (DDDA, or 3DA). Embodiments of the present disclosure are particularly applied in the lending industry, but in other examples, may be applicable to other financial industries. For example, features of the present disclosure use data to automate decisions relating to commercial loans, mortgage loans, home equity loans and so forth. Typically, the decisions occur during a loan application process. A loan application process consists of multiple workplans that involve multiple parties, culminating in either an acceptance or a rejection by the lender of the loan application. If the loan application is accepted, the lender is in a position to issue the loan to the customer.” (Paramasamy, (Col. 3: lines 54-67).
Regarding Claim 2:
James teaches:
wherein the first machine learning model is a trained classifier network. (James, classification algorithms may include support vector machines, discriminant analysis, naive Bayes, and nearest neighbor algorithms that be used for supervised learning. Similarly, various regression algorithms such as linear regression, GLM, SVR, GPR, ensemble methods, decision trees, and neural networks may be used for supervised learning….to assess model performance the data can be partitioned into a training set and a validation or test set. Training set used to construct the classifiers and the test set is used to assess their performance. (See, Para. 49, 50)).
Regarding Claims 3 and 9:
James teaches:
wherein the second machine learning model is trained using a regression algorithm. (James, Gradient boosting is a machine learning technique for regression and classification problems, which produces a prediction model in the form of an ensemble of weak prediction models, typically decision trees. …..ML model B 305 may use a logistic regression algorithm. Logistic regression is a classification algorithm used to assign observations to a discrete set of classes. Unlike linear regression which outputs continuous number values, logistic regression transforms its output using the logistic sigmoid function to return a probability value which can then be mapped to two or more discrete classes. ML model n 307 may use other algorithms. (See, Para. 35, 47)).
Regarding Claims 4 and 10:
James does not specifically teach that the data acquisition engine is further configured to: retrieve one or more compliance rules; and transform the validated data to enforce compliance with the one or more compliance rules.
However, Paramasamy teaches the following limitations:
wherein the data acquisition engine is further configured to: retrieve one or more compliance rules; and transform the validated data to enforce compliance with the one or more compliance rules. (Paramasamy, An example of using data element groupings, a document context grouping of data elements relating to mortgage insurance can be assigned the compliance category and thereby easily matched with the corresponding mortgage 25 insurance compliance requirements to determine if the requirements have been met. For instance, it is a compliance requirement that every loan should have mortgage insurance and for the lender to disclose in the Good Faith Estimate or Loan Estimate document provided to the customer the existence of any mortgage insurance. When a mortgage insurance document is received, data elements are acquired from the mortgage insurance document and assigned to the compliance category, (See, Col. 13: lines 11-63; Col. 7: lines 61-67; Col. 8: line 1));
and transform the validated data to enforce compliance with the one or more compliance rules. (Paramasamy, When a mortgage insurance document is received, data elements are acquired from the mortgage insurance document and assigned to the compliance category, allowing the system to quickly ascertain if the relevant compliance requirements have been met, and then automatically performing an action ( completing workplan, issuing a notification of incomplete information, etc.) based on the evaluation. At the modeling phase 306, segments of prepared data elements of the acquired data are created building clusters of elements in intelligent ways for loan application workplans using context modeling algorithms 216 and incorporating and positioning the newly acquired data elements relative to those clusters to create segments. (See, Col. 13: lines 11-63; Col. 7: lines 61-67; Col. 8: line 1)).
It would have been obvious to one of ordinary skill in the art before the effective filing of the claimed invention to have modified James with the features of Paramasamy’s system because “the present disclosure is directed to systems and methods that employ data driven decision automation (DDDA, or 3DA). Embodiments of the present disclosure are particularly applied in the lending industry, but in other examples, may be applicable to other financial industries. For example, features of the present disclosure use data to automate decisions relating to commercial loans, mortgage loans, home equity loans and so forth. Typically, the decisions occur during a loan application process. A loan application process consists of multiple workplans that involve multiple parties, culminating in either an acceptance or a rejection by the lender of the loan application. If the loan application is accepted, the lender is in a position to issue the loan to the customer.” (Paramasamy, (Col. 3: lines 54-67).
Regarding Claims 5 and 11:
James does not specifically teach that the borrower profile comprises one or more access rules define one or more lender institutions which the borrower has authorized to the data in the borrower profile.
However, Paramasamy teaches the following limitation:
wherein the borrower profile comprises one or more access rules define one or more lender institutions which the borrower has authorized to the data in the borrower profile. (Paramasamy, the interface 800 also includes a loan information pane 804 that provides a pictorial representation of the status of 25 the transaction. In this instance, the credit profile for the borrower is shown, including credit, income, assets, etc. The pane 804 includes example circle and line bars to illustrate the relative completeness of each depicted item associated with the customer's credit profile….as the information about the assets and other conditions change over time, the quantification can increase or decrease, with a 100 percent qualification profile indicating that the customer is approved for the loan transaction. It is appreciated that pane 804 is customized specifically to the customer, and the information shown and presented will dynamically change based on the particular customer; The one or more server computers can also obtain, via a network, data from a number of different third-party databases. Such databases can include, for example, government databases that store taxpayer information, statutory and regulatory information, zoning information, property lien information, survey and title information, homeowners' association information, and so forth. Other databases can include those of credit rating associations, real estate organizations, financial aggregator organizations, other financial institutions, insurance providers, etc. Pre-authorization may be needed, e.g., from the customer, before the lender is granted access to information related to the customer's loan application from one or more of these databases (Col.7: lines 29-67; Col. 9: lines 9-45; Col. 23: lines 24-43)).
It would have been obvious to one of ordinary skill in the art before the effective filing of the claimed invention to have modified James with the features of Paramasamy’s system because “the present disclosure is directed to systems and methods that employ data driven decision automation (DDDA, or 3DA). Embodiments of the present disclosure are particularly applied in the lending industry, but in other examples, may be applicable to other financial industries. For example, features of the present disclosure use data to automate decisions relating to commercial loans, mortgage loans, home equity loans and so forth. Typically, the decisions occur during a loan application process. A loan application process consists of multiple workplans that involve multiple parties, culminating in either an acceptance or a rejection by the lender of the loan application. If the loan application is accepted, the lender is in a position to issue the loan to the customer.” (Paramasamy, (Col. 3: lines 54-67).
Regarding Claims 6 and 12:
James does not specifically teach that an application programming interface comprising a second plurality of programming instructions stored in the memory which, when operating on the processor, causes the computing device to: transmit the validated data in the borrower profile to a loan origination system associated with the one or more authorized lender institutions.
However, Paramasamy teaches the following limitation:
further comprising an application programming interface comprising a second plurality of programming instructions stored in the memory which, when operating on the processor, causes the computing device to: transmit the validated data in the borrower profile to a loan origination system associated with the one or more authorized lender institutions.. (Paramasamy, the interface 800 also includes a loan information pane 804 that provides a pictorial representation of the status of 25 the transaction. In this instance, the credit profile for the borrower is shown, including credit, income, assets, etc. The pane 804 includes example circle and line bars to illustrate the relative completeness of each depicted item associated with the customer's credit profile….as the information about the assets and other conditions change over time, the quantification can increase or decrease, with a 100 percent qualification profile indicating that the customer is approved for the loan transaction. It is appreciated that pane 804 is customized specifically to the customer, and the information shown and presented will dynamically change based on the particular customer; The one or more server computers can also obtain, via a network, data from a number of different third-party databases. Such databases can include, for example, government databases that store taxpayer information, statutory and regulatory information, zoning information, property lien information, survey and title information, homeowners' association information, and so forth. Other databases can include those of credit rating associations, real estate organizations, financial aggregator organizations, other financial institutions, insurance providers, etc. Pre-authorization may be needed, e.g., from the customer, before the lender is granted access to information related to the customer's loan application from one or more of these databases (Col.7: lines 29-67; Col. 9: lines 9-45; Col. 23: lines 24-43)).
It would have been obvious to one of ordinary skill in the art before the effective filing of the claimed invention to have modified James with the features of Paramasamy’s system because “the present disclosure is directed to systems and methods that employ data driven decision automation (DDDA, or 3DA). Embodiments of the present disclosure are particularly applied in the lending industry, but in other examples, may be applicable to other financial industries. For example, features of the present disclosure use data to automate decisions relating to commercial loans, mortgage loans, home equity loans and so forth. Typically, the decisions occur during a loan application process. A loan application process consists of multiple workplans that involve multiple parties, culminating in either an acceptance or a rejection by the lender of the loan application. If the loan application is accepted, the lender is in a position to issue the loan to the customer.” (Paramasamy, (Col. 3: lines 54-67).
Regarding Claim 8:
James teaches:
wherein the plurality of convolutional layers includes three layers for three different granularities. (James, the relevance of the weighted features (Objective function 301) would be tested by the machine learning model development module 115 using machine learning models (e.g. ML model A 303, ML model B 305 and ML model n 307 implemented by the machine learning module 105. the results of such tests are used to train the machine learning module 105. For example, when input is received from machine learning processing subsystem 107 (Preprocessing and feature engineering), ML model A 303 may use a gradient boosting machine algorithm…. regression algorithms such as …. neural networks may be used for supervised learning. (See, Para. 47, 49)); normalize documents of varying dimensions using adaptive pooling; (James, Machine learning module 105 pre-processes the data in pre-processing subsystem 109. Pre-processing subsystem 109 pre-processes the data by formatting, cleaning, and sampling the data. The formatting step converts the data into a format that is suitable for use by the machine learning module 105….Finally standardizing numeric features (scaling values to lie within the same range) is necessary in regularized logistic regression algorithms, and the automated feature engineering subsystem 111 adds the steps because it knows that this step constitutes best practice. (See, Para. 32-35; Fig. 1));
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure is the following:
Fidanza (U.S. Patent Application Pub. No. US 2020/0349641 A1) “System and method for determining credit and issuing a business loan using tokens and machine learning”
Iyer (U.S. Patent Application Pub. No. US 2022/0301031 A1) “Machine learning based automated product classification”
Jeong (U.S. Patent Application Pub. No. US 2022/0076081 A1) “Modular machine learning systems and methods”
Campos (U.S. Patent Application Pub. No. US 2023/0097897A1) “Automated Model Selection”
Rezvani (U.S. Patent Application Pub. No. US 2021/0064866) “Automatic document classification using machine learning”
Mazor (U.S. Patent Application Pub. No. US 2022/0198316) “Systems and Methods for Automatic Extraction of Classification Training Data”
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MOHAMMED H MUSTAFA whose telephone number is (571)270-7978. The examiner can normally be reached M-F 8:00 - 5:00.
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, MICHAEL W. ANDERSON can be reached on 571-270-0508. 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.
/MOHAMMED H MUSTAFA/Examiner, Art Unit 3698/Mike Anderson/Supervisory Patent Examiner, Art Unit 3693
Prosecution Insights