Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
This Non-Final Rejection is filed in response to Applicant Arguments/Remarks Made in an Amendment filed 05/23/2025.
Claims 1, 9, & 15 are amended.
In light of the amendments the U.S.C. 112f rejections are respectfully withdrawn.
Claims 1-5 & 8-18 remain pending.
Response to Arguments
Argument 1, Applicant argues in Applicant Arguments/Remarks Made in an Amendment filed 05/23/2025, pg. 7-10 that “the alleged judicial exception in integrated in to a practical application of an improved machine learning model generation, which weights financial data based on risk, and generates a machine learning model based on the weighted financial data” and thus overcomes the 101 rejection.
Response to Argument 1, the examiner respectfully disagrees. Under a broadest reasonable interpretation, These steps fall under the abstract of idea of mental processes that can be performed in the human mind, or by a human using a pen a paper. A data scientist could be supervising the machine learning model generation and use their own evaluation to adjust the weights according to their own discretion. These components are recited at a high level of generality and amount to no more than instructions to implement the abstract idea using generic machine learning components, or merely uses a machine learning components as a tool to perform an abstract idea. While applicant cites reasons as an improvement to classification of transaction classification methods, simply weighting certain values more than others does not represent an improvement to the technology or technical field of machine learning, but rather represents a human directing a machine learning model to focus on certain data of their own evaluation.
Argument 2, Applicant argues in Applicant Arguments/Remarks Made in an Amendment filed 05/23/2025, pg. 11 that Verma does not teach, “generating a model for detecting illegal financial transaction data, on a basis of the financial transaction data and the weight set to the financial transaction data such that data associated with a larger weight has a greater impact on the generated model than data with a lower weight”.
Response to Argument 1, the examiner respectfully disagrees. Verma describes using machine learning as part of an anomaly detection progress that classifies a whole user profile according to their own weighted transactions. Wherein user transactions are given larger or smaller weights that contribute to an over risk score according to how much larger a dollar amount spent is over a defined average spend limit. Thus Verma teaches that a transaction has a much higher than average dollar spend is weighted with a higher likelihood of fraud than a lower dollar amount that is closer to an average spend. Thus the anomaly detection model generated for a user will assign a higher overall risk of fraud to a user, impacted by how they may have had a high risk transaction. The following paragraphs of Verma provide evidence of this interpretation.
para. [0054], Various machine learning and/or anomaly detection process can be performed on historical expense data associated with an employee to identify habitual, opportunistic and/or accidental fraud in employee spending, Quantitative and/or qualitative features can be measured against other employee cohorts in the enterprise… example, qualitative features can include job function and/or rank within company.
para. [0055-0059], Use machine learning to classify whether a report submitted by an employee is of high, medium or low risk by assigning weights to different policy violations across expense lines and studying employee profile data… For every user in a company, risk and behavior features are constructed based on the above and fed into a machine learning algorithm such as SVM linear kernel model for predicting an overall risk score for the user that may be HIGH, MEDIUM or LOW.
Claim Rejections - 35 USC § 101
Step 1: Claim 1 is an apparatus type claim. The claim recites at least one step or act of determining classifying a transaction, setting weights, and generating model based on classification and set weights. Claims 9 & 15 recite similar limitations. Therefore, claims 1-5 and 8-18 are directed to either a process, machine, manufacture or composition of matter.
With respect to claim 1:
An information processing apparatus comprising:
a memory in which processing instructions are stored; and
at least one processor configured to execute the processing instructions to perform processing of:
(B) setting a classification of financial transaction data of an illegal transaction out of financial transaction data that is a learning target, on a basis of a value of a specific amount of money which is an attribute constituting the financial transaction data, wherein the classification comprises a weight having a value that increases as the value of the specific amount of money increases, wherein the weight represents a risk of illegality; and
(C) generating a model for detecting illegal financial transaction data, on a basis of the financial transaction data and the weight set to the financial transaction data, such that data associated with a larger weight has a greater impact on the generated model than data with a lower weight.
2A Prong 1:
Step (B) recites for classifying an illegal financial transaction by weighting a financial transaction. These limitations are directed to an abstract idea in which a human may look at data from a model, write down observations on said data, and use written down observations to assign larger weights as the data observed i larger than a certain value. Under a broadest reasonable interpretation, These steps fall under the abstract of idea of mental processes that can be performed in the human mind, or by a human using a pen a paper, wherein the training and generation of a model merely involve ingestion of the human evaluated data by an off the shelf machine learning model.
Thus the broadest reasonable interpretation of steps (B) is that those steps fall within the mental process groupings of abstract ideas because they cover concepts performed in the human mind, including observation, evaluation, judgment, and opinion. See MPEP 2106.04(a)(2), subsection III.
2A Prong 2: This judicial exception is not integrated into a practical application.
Additional elements:
The limitations (A & C), “a memory in which processing instructions are stored; and at least one processor configured to execute the processing instructions” & “generating a model for detecting illegal financial transaction data, on a basis of the financial transaction data and the weight set to the financial transaction data, such that data associated with a larger weight has a greater impact on the generated model than data with a lower weight”. These components are recited at a high level of generality and amount to no more than instructions to implement the abstract idea using generic computer components, or merely uses a computer as a tool to perform an abstract idea. Further, the model generated is an off the shelf model, which amounts to no more than an “apply it” of the abstract idea (MPEP 2106.05(f)).
2B: Even when considered in combination, these additional elements represent mere instructions to implement an abstract idea or other exception on a computer and insignificant extra-solution activity, which do not provide an inventive concept.
Claims 9 & 15 are the method and medium claims reciting similar limitations to Claim 1. The same arguments as stated above apply, because they cover the same mental functions recited in Claim 1 and is rejected as ineligible subject matter under 35 U.S.C. 101 for the same reasons as stated in claim 1.
As per Claim 2, the claim includes additional mental steps similar to Claim 1, and therefore will be rejected for similar reasons. The limitation of, “wherein the at least one processor sets the classification of the financial transaction data on a basis of a value of an attribute constituting the financial transaction data”, are elements related to mere data gathering and thus are insignificant extra-solution activity. Thus Claim 2 is directed to an abstract idea.
As per Claim 3, the claim includes additional mental steps similar to Claim 1, and therefore will be rejected for similar reasons. The limitation of, “wherein the at least one processor sets a weight having a predetermined value as the classification of the financial transaction data, and generates the model on a basis of the financial transaction data and the weight set to the financial transaction data”, are elements related to mere data gathering and thus are insignificant extra-solution activity. Thus Claim 3 is directed to an abstract idea.
As per Claim 4 the claim includes additional mental steps similar to Claim 3, and therefore will be rejected for similar reasons. The limitation of, “wherein the at least one processor sets, to the financial transaction data, the weight having a value that is larger as a degree of significance is higher, the degree of significance being determined based on a preset criterion according to a value of an attribute constituting the financial transaction data”, are elements related to “mental process” groupings because it requires an assessment of data information in that one larger value is assigned a larger weight and applying the exception using generic computer components, which do not integrate the recited judicial exception into practical application. Moreover, the recited determination is simple enough that it can be practically performed in the human mind. Even if most humans would use a physical aid, like a pen and paper or a calculator, to make such calculations, the use of a physical aid would not negate the mental nature of this limitation. See MPEP 2106.04(a)(2), subsection III.B.. Thus Claim is directed to an abstract idea.
As per Claim 5, the claim includes additional mental steps similar to Claim 1, and therefore will be rejected for similar reasons. The limitations of, “wherein the at least one processor :stores classification data representing a correspondence relationship between a value of an attribute constituting the financial transaction data and the classification; and generates new classification data and updates the classification data, on a basis of the value of the attribute constituting the financial transaction data”, are elements related to mere data gathering and output recited at a high level of generality, applying the exception using generic computer components, and thus are insignificant extra-solution activity. Thus Claim 5 is directed to an abstract idea.
As per Claim 8, the claim includes additional mental steps similar to Claim 1, and therefore will be rejected for similar reasons. The limitation of, “wherein the at least one processor detects the illegal financial transaction data with use of the model, from detection target financial transaction data”, are elements related applying the exception using generic computer components, and thus are insignificant extra-solution activity. Thus Claim 8 is directed to an abstract idea.
As per Claim 17, the claim includes additional mental steps similar to Claim 1, and therefore will be rejected for similar reasons. The limitation of, “wherein the at least one processor sets the weight, the weight having a value that is larger as a value of an amount of money of damage of the financial transaction data of the illegal transaction involving the damage is larger”, are elements related to “mental process” groupings because it requires an assessment of data information in that one larger value is assigned a larger weight and applying the exception using generic computer components, which do not integrate the recited judicial exception into practical application. Thus Claim 17 is directed to an abstract idea.
As per Claim 17, the claim includes additional mental steps similar to Claim 1, and therefore will be rejected for similar reasons. The limitation of, “wherein the at least one processor sets the weight according to values of a plurality of attributes including a value of another attribute which is different from the amount of money, and constitutes the financial transaction data”, are elements related to mere data gathering and thus are insignificant extra-solution activity and applying the exception using generic computer components, which do not integrate the recited judicial exception into practical application. Thus Claim 17 is directed to an abstract idea.
Claims 10-14 & 16 are respectively the method and apparatus claims reciting similar limitations to Claims 2-8 and is rejected for similar reasons.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1-5, 8-18 is/are rejected under 35 U.S.C. 102(a)(2) as being anticipated by U.S. Patent Application Publication. NO. 20160358268 “Verma”.
Claim 1:
Verma teaches an information processing apparatus comprising: a memory in which processing instructions are stored; and at least one processor configured to execute the processing instructions to perform processing instructions (i.e. para. [0071], Fig. 1, computing system 1100 may include, for example, a processor, memory) of: when setting a classification of financial transaction data of an illegal transaction on out of financial transaction data that is a learning target, on a basis of a value of a specific amount of money which is an attribute constituting the financial transaction data (i.e. para. [0041], “Machine learning can be used to classify an expense as opportunistic or accidental fraud based on employee's history…Machine learning can also be used for spend profiling. For example, it can be determined if spending one-thousand dollars ($1000.00) on a customer meal is considered normal for that company ”, wherein the BRI for classifying encompasses how a ML system may set a classification for a target expense data. Wherein the classification is the chance that the expense is fraudulent or not based on if the specific amount of money representing expense data satisfies fraudulent condition such as the value of the specific expense amount being above a company role limit), wherein the classification comprises a weight having a value that increases as the value of the specific amount of money increases (i.e. para. [0061], “K-nearest neighbor clustering algorithms can be used to detect anomalies. For example, process 400 can construct the distance of the current expense to its k.sup.th nearest neighbor. If the distance is high comparing to its similar purchase, then the expense is flagged as high risk”, wherein a transaction may be classified as a “high” risk when the specific amount of money is a substantial increase (such as a standard deviation away) from an average spend. Thus as the amount of money spent increases above a policy average, the transaction is classified a higher risk weight over a lower risk weight) wherein the weight represents a risk of illegality (i.e. para. [0045, 0055], “a company can have an annual limit for tuition reimbursement at two-thousand dollars ($2000.00). It can be determined if the employee is claiming great than said limit… machine learning to classify whether a report submitted by an employee is of high, medium or low risk by assigning weights to different policy violations across expense lines”, wherein a transaction may be weighted to have a larger value for probability of fraud as the amount of money that constitutes the transaction data is larger than a given limit of a policy. Wherein the fraud detection model may generate a risk score on a basis that the transaction data and the weighting of the policy violation); and
generating a model for detecting illegal financial transaction data, on a basis of the financial transaction data and the weight set to the financial transaction data (i.e. para. [0054], “Various machine learning and/or anomaly detection process can be performed on historical expense data associated with an employee to identify habitual, opportunistic and/or accidental fraud in employee spending, Quantitative and/or qualitative features can be measured against other employee cohorts in the enterprise… example, qualitative features can include job function and/or rank within company”, wherein a trained model for detecting fraudulent accounts may be generated on the classifications set for a company considers fraud) such that data associated with a larger weight has a greater impact on the generated model than data with a lower weight (i.e. para. [0055-0059], “Use machine learning to classify whether a report submitted by an employee is of high, medium or low risk by assigning weights to different policy violations across expense lines and studying employee profile data… For every user in a company, risk and behavior features are constructed based on the above and fed into a machine learning algorithm such as SVM linear kernel model for predicting an overall risk score for the user that may be HIGH, MEDIUM or LOW”, wherein a transaction flagged with a high risk has a greater impact than a transaction flagged with a low risk. The examiner notes that the BRI for impact is quite broad and encompasses metrics such as how transactions many standard deviations amount of dollars over an average dollar spend have a larger weight for impacting the predictive model predicting a high risk of fraud for a user than transactions weighted with a lower amount of dollars over an average dollar spend).
Claim 2:
Verma teaches the information processing apparatus according to claim 1.
Verma further teachers wherein the at least one processor
sets the classification of the financial transaction data on a basis of a value of an attribute constituting the financial transaction data (i.e. para. [0045]. “a company can have an annual limit for tuition reimbursement at two-thousand dollars ($2000.00). It can be determined if the employee is claiming greater than said limit”, wherein the process may classify an expense on the basis that it contains a transaction greater than a company limit)
Claim 3:
Verma teaches the information processing apparatus according to claim 1.
Verma further teachers wherein the at least one processor
sets a weight having a predetermined value as the classification of the financial transaction data, and generates the model on a basis of the target data and the weight set to the financial transaction data (i.e. para. [0055], “an expense-report risk profile can be determined. Use machine learning to classify whether a report submitted by an employee is of high, medium or low risk by assigning weights to different policy violations across expense lines and studying employee profile data. The system is by using historical risk and behavioral data for employees”, wherein the BRI for setting a weight having a predetermined value as the classification encompasses how the machine learning model may use predetermined and historical data to weight policy violations to generate a risk model for fraud).
Claim 4:
Verma teaches the information processing apparatus according to claim 3.
Verma further teachers wherein the at least one processor
sets, to the financial transaction data, the weight having a value that is larger as a degree of significance is higher, the degree of significance being determined based on a preset criterion according to a value of an attribute constituting the financial transaction data (i.e. para. [0055], an expense-report risk profile can be determined. Use machine learning to classify whether a report submitted by an employee is of high, medium or low risk by assigning weights to different policy violations across expense lines”, wherein a model sets a weight that determines if transaction has a value of low-high risk of fraud based on an attribute of if the transaction violated a policy limit).
Claim 5:
Verma teaches the information processing apparatus according to claim 1.
Verma further teachers wherein the at least one processor: stores classification data representing a correspondence relationship between a value of an attribute constituting the financial transaction data and the classification (i.e. para. [0045], travel guidelines can be maintained in a database and compared with the employee's claim travel in an expense report); and generates new classification data and updating the classification data, on a basis of the value of the attribute constituting the financial transaction data (i.e. para. [0054], “process 100 can obtain enrichment/augmentation of expense data with web scale data. Step 106 can use web-scale sources (e.g. web search, yelp®, TripAdvisor®, Wikipedia, DBPedia, etc.) to augment user-entered data”, wherein it is noted that new classification data may be generated in that classification data from the web may be used to augment and update the expense data with multiple other sources, on the basis if the data indicates that a user is committing fraud.).
Claim 8:
Verma teaches the information processing apparatus according to any of claim 1.
Verma further teachers wherein the at least one processor
Detects the illegal financial transaction data with use of the model, from a detection target financial transaction data (i.e. para. [0054], FIG. 4 illustrates an example process 400 that uses machine learning to detect anomalies in expense report data).
Claim 9:
Claim 9 is the method claim reciting similar limitations to Claim 1 and is rejected for similar reasons.
Claim 10:
Claim 10 is the method claim reciting similar limitations to Claim 2 and is rejected for similar reasons.
Claim 11:
Claim 11 is the method claim reciting similar limitations to Claim 3 and is rejected for similar reasons.
Claim 12:
Claim 12 is the method claim reciting similar limitations to Claim 4 and is rejected for similar reasons.
Claim 13:
Claim 13 is the method claim reciting similar limitations to Claim 5 and is rejected for similar reasons.
Claim 14:
Claim 14 is the method claim reciting similar limitations to Claim 8 and is rejected for similar reasons.
Claim 15:
Claim 15 is the apparatus claim reciting similar limitations to Claim 1 and is rejected for similar reasons.
Claim 16:
Claim 16 is the apparatus claim reciting similar limitations to Claim 8 and is rejected for similar reasons.
Claim 17:
Verma teaches the information processing apparatus according to claim 1.
Verma further teachers wherein the at least one processor sets the weight, the weight having a value that is larger as a value of an amount of money of damage of the financial transaction data of the illegal transaction involving the damage is larger (i.e. para. [0045, 0055], “a company can have an annual limit for tuition reimbursement at two-thousand dollars ($2000.00). It can be determined if the employee is claiming great[er] than said limit… machine learning to classify whether a report submitted by an employee is of high, medium or low risk by assigning weights to different policy violations across expense lines”, wherein a transaction may be weighted to have a larger probability of fraud as the amount of damage that constitutes the potentially fraudulent transaction data is larger than a given limit of a policy. Wherein the fraud detection model may generate a risk score on a basis that the transaction data and the weighting of the policy violation).
Claim 18:
Verma teaches the
Verma further teachers the information processing apparatus according to claim 1, wherein the at least one processor sets the weight according to values of a plurality of attributes including a value of another attribute which is different from the amount of money, and constitutes the financial transaction data (i.e. para. [0045], “For example, employee's purchase of alcohol can be determined to be or not be in compliance with a company policy. The location of the expense/purchase can also be identified. It can be determined if the location is allowed by company policy. For example, it can be determined if loess bar is a strip club or a restaurant”, wherein a plurality of values of other attributes such as location and transaction compliance may be weighted if the value of the attributes violates the company policy).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
U.S. Patent Application Publication. NO. 20110191225 “Mark”, in para. [0093], that It would be important know which files had the greatest equivalent US dollar amount, in order to make sure the systems processing the transactions with the greatest equivalent US dollar amount were made a priority over the other transactions.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DAVID H TAN whose telephone number is (571)272-7433. The examiner can normally be reached M-F 7:30-4:30.
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, Cesar Paula can be reached on (571) 272-4128. 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.
/D.T./Examiner, Art Unit 2145
/CESAR B PAULA/Supervisory Patent Examiner, Art Unit 2145