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 .
Acknowledgments
The RCE filed on 03/23/26 is acknowledged.
Status of Claims
Claims 1-3, 5, 7, 10-13, 17, 18 and 21 are pending.
In the Amendment filed on 02/27/26 (entered via the RCE filed on 03/23/26), claims 1, 2, 10, 11 and 17 were amended, claims 9, 16 and 20 were cancelled, and no claims were added. (Claims 4, 6, 8, 14, 15, 19 and 22 were cancelled in one or more previous papers.)
Claims 1-3, 5, 7, 10-13, 17, 18 and 21 are rejected.
Response to Arguments
Regarding the rejection under 35 U.S.C. 101
Applicant's arguments have been fully considered but are not persuasive.
The Office responds to Applicant arguments below, under the headings used by Applicant in the instant Response. In the below discussion, page numbers refer to Applicant's instant Response, unless indicated otherwise.
A. Step 2A – Prong Two Analysis (pp. 2-5)
Applicant argues:
Applicant submits that the claims as amended are integrated into a practical application that addresses specific technological problems in computerized payment processing systems. The claims recite a comprehensive technological solution for real-time transaction processing that goes beyond any alleged abstract idea. Specifically, amended claim 1 recites
training a classification machine learning model on transaction outcomes of prior transactions;
training a causal inference machine learning model on the transaction outcomes of the prior transactions;
classifying the transaction using the trained classification machine learning model trained to predict a probability of each of one or more possible outcomes of transactions, wherein each possible outcome is an end result of processing the transaction;
configuring the classification machine learning model is configured to compute a probability score vector for the transaction, wherein the probability score vector indicates a probability for each of the one or more possible outcomes of the transaction;
analyzing the transaction by the trained causal inference machine learning model, including computing one or more probability mass vectors for the transaction, wherein:
…
selecting a treatment to be applied to the transaction by combining an output of the classification machine learning model and an output of the causal inference machine learning model, wherein the selected treatment is based on the probability score vector and the one or more probability mass vectors;
…
feeding back the transaction outcome to the classification machine learning model and the causal inference machine learning model to further train the classification machine learning model and the causal inference machine learning model.
This recitation describes a specific technological implementation that manipulates the structure and content of digital transaction data in real-time to optimize transaction processing outcomes and utilizes the output of two distinct machine learning models working in tandem to produce the result.
The technological nature of this solution becomes more apparent when considering the specific data manipulations recited and the interactions between the two machine learning models. The claims specify that treatments include “removing a data field, editing a data field, adding a data field, or reordering the data fields in the transaction.” The claims further specify that the output of the classification machine learning model and the output of the causal inference machine learning model are combined to select a treatment to be applied to the transaction. Furthermore, the transaction outcome is fed back to the classification machine learning model and the causal inference machine learning model for further training of the machine learning models. (Remarks, pp. 2-31; underlining added)
The Office responds:
The manipulation of data (performed here in the service of efficiently processing transactions while reducing transaction fraud) is an abstract idea, and is not technological. The machine learning limitations amount merely to training (including retraining through feedback) and using machine learning models to apply the abstract idea. The machine learning models and their training and use are recited at a high level of generality and are not described. As such, they do not reflect any improvement in computer functioning or other technology, but are merely used off-the-shelf in their ordinary capacities (as tools), applied after the fact to the abstract idea. MPEP 2106.05(f). In particular, contrary to Applicant's argument, the claims do not in fact recite any significant interaction or working in tandem of the two machine learning models. Rather, each machine learning model is used separately to yield different outputs, and the outputs are merely "combined" in a manner that is not described in the claims at all.2
Applicant further argues:
These operations are inherently technological, as they involve the programmatic modification of digital data structures during transaction processing. In the context of high-speed electronic payment processing, transactions must be analyzed and modified within milliseconds to maintain system performance. (Remarks, pp. 3-4)
The Office responds:
Applicant's argument describes the use of generic computer elements ("programmatic"; "digital") to apply an abstract idea ("modification of … data structures during transaction processing"). The speed of analysis merely reflects the fact that the analysis is automated by use of computers, i.e., generic computer elements. Thus, the expression "inherently technological" refers merely to automation by generic computers.
Applicant further argues:
Furthermore, the claims integrate machine learning technology in a specific and non-abstract manner. The classification machine learning model is “trained on transaction outcomes of prior transactions” and is “configured to compute a probability score vector for the transaction.” The causal inference machine learning model computes “probability mass vectors” that capture “impact values and associated probabilities” of different treatments. This represents a sophisticated technological implementation that uses historical transaction data to train two distinct machine learning models, which then generate probability distributions for real-time decision making. The feedback mechanism described in the claims, where “the transaction outcome” is fed back “to the classification machine learning model and the causal inference machine learning model to further train the classification machine learning model and the causal inference machine learning model,” creates a closed-loop learning system that continuously improves its performance based on actual transaction results. (Response, p. 4)
The Office responds:
Again, this description reiterates the separate, independent use of the two machine learning models, which is not interactive in any technologically substantive way.3 Again, the description reiterates that the machine learning limitations amount merely to training (including retraining through feedback) and using machine learning models to apply the abstract idea; they are recited at a high level of generality and are not described; they are merely used off-the-shelf in their ordinary capacities (as tools), applied after the fact to the abstract idea. As such, they do not reflect any improvement in computer functioning or other technology.
Applicant further argues:
The claims address the technical challenge of making real-time modifications to transaction data while maintaining electronic payment processing system performance and reliability. As described in the specification, the system enables “in-flight adjustments to the transaction’s data fields” (Applicant’s specification as published, paragraph 0009) to “complete the transaction approval process as quickly as possible” (Id., paragraph 0005). This technical problem is unique to computerized transaction processing systems, where the speed and accuracy of data manipulation directly impact system performance and user experience. Indeed, outside of the realm of a computerized transaction processing system, there would be no need to analyze or adjust “data fields” included in a “transaction” because such fields do not exist nor are they necessary for enabling a non-computerized transaction to be approved. (Response, pp. 4-5)
The Office responds:
Again, Applicant's argument reflects the fact that the abstract idea is applied using generic computer technology in its ordinary capacity / recited at a high level of generality. As such, Applicant's claimed invention merely exploits the ordinary advantages of such technology, such as speed and accuracy of data manipulation, for the benefit of the abstract idea.
As for Applicant's apparent assertion that its claims could not exist outside the computer realm, the Office responds that (1) the problem(s) addressed by Applicant's claims (e.g., preventing fraud, preventing drop-offs, speeding up transaction processing, deciding whether to require heightened authentication, or the like; see Applicant's specification, 0005-0010) do not arise specifically from computer/Internet technology, but rather from a business process, and (2) the solution constituted/provided by Applicant's claims does not improve upon the pre-existing/underlying computer/other technology, but merely uses that pre-existing/underlying technology off-the-shelf as a tool in its ordinary capacity, and rather improves only upon the abstract idea, if it improves anything at all. As such, the instant claims are not sufficiently analogous to those of decisions like DDR such as would render the claims eligible under 35 U.S.C. 101.
Applicant further argues:
The integration of multiple analytical machine learning models working in concert further demonstrates the technological nature of the solution recited in the claims. The claims require both a classification model that generates probability score vectors and a separate analysis by a causal inference model that computes probability mass vectors, with the treatment selection based on both types of analysis. This multiple machine learning model approach represents a sophisticated technological architecture that coordinates different analytical engines to optimize transaction processing decisions. Using the separate classification machine learning model and causal inference machine learning model permits finer-grained decision making (see Applicant’s specification as published, paragraph 0067) and provides more transparency and explainability of the decision to apply treatments (see Id., paragraph 0068). (Response, p. 5; underlining added)
The Office responds:
Applicant's argument highlights the fact that the two machine learning models operate entirely separately and independently of each other, and do not interact in any technologically substantive way.4 The putative improvement is a putative improvement in the abstract idea, namely, the use of additional data analyses that allegedly provides "finer-grained decision making[,] …more transparency and explainability." These data analyses are automated by a generic computer, which as such serves merely to apply the abstract idea.
In sum, as per the discussion above, Applicant's claimed subject matter amounts to an abstract idea applied using generic computer elements. As such, the abstract idea is not integrated into a practical application.
B. Step 2B Analysis (pp. 5-8)
Applicant argues that the claims "amount to 'significantly more' than any asserted abstract idea." Response, p. 5
The Examiner respectfully disagrees.
Applicant argues on the basis of the same claimed subject matter as was argued in respect of Step 2A. This claimed subject matter has been addressed and analyzed above. It amounts merely to an abstract idea applied using generic computer elements. The alleged improvement as described amounts merely to a narrowing of the abstract idea, which does not render a claim eligible under 35 U.S.C. 101.
It is also noted that the rejection did not / does not assert that the claimed additional elements are well understood, routine, and conventional. To the extent that Applicant argues this point (see, e.g., "non-conventional," p. 7), the argument is moot. Rather, in step 2B, the rejection asserted / asserts that the additional elements, both individually and in combination, amount to no more than generic computer tools to perform the abstract idea. As stated in Alice, “[s]imply appending conventional steps, specified at a high level of generality,’ [is] not ‘enough’ to supply an ‘inventive concept.’ (134 S. Ct. at 2357 (citing Mayo, 132 S. Ct. at 1300, 1297, 1294). This statement was making clear that limiting the use of the abstract idea by recitation of generic computer hardware is insufficient to establish eligibility. See MPEP 2106.05 I.A. ("Limitations that the courts have found not to be enough to qualify as "significantly more" when recited in a claim with a judicial exception include: i. Adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, e.g., a limitation indicating that a particular function such as creating and maintaining electronic records is performed by a computer, as discussed in Alice Corp., 573 U.S. at 225-26, 110 USPQ2d at 1984 (see MPEP § 2106.05(f))"); MPEP 2106.05(f) ("As explained by the Supreme Court, in order to make a claim directed to a judicial exception patent-eligible, the additional element or combination of elements must do "‘more than simply stat[e] the [judicial exception] while adding the words ‘apply it’". Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, for example, claims that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not render an abstract idea eligible. Alice Corp., 573 U.S. at 223, 110 USPQ2d at 1983. See also 573 U.S. at 224, 110 USPQ2d at 1984 (warning against a § 101 analysis that turns on "the draftsman’s art").")
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 1-3, 5, 7, 10-13, 17-18 and 21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Claims 1-3, 5, 7, 10-13, 17-18 and 21 are directed to a method, system, or transaction analysis unit, which are/is one of the statutory categories of invention. (Step 1: YES)
Claims 1, 10 and 17 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claims recite a method, system, and transaction analysis unit for payment processing including applying a treatment to the transaction based on analysis of the transaction (in order to prevent fraud while avoiding drop-offs, as per specification 0006-0010). For claims 1, 10 and 17 (claim 1 being deemed representative), the limitations (indicated below in bold) of:
receiving a transaction, wherein the transaction includes one or more data fields;
training a classification machine learning model on transaction outcomes of prior transactions;
training a causal inference machine learning model on the transaction outcomes of the prior transactions;
classifying the transaction using the trained classification machine learning model to predict a probability of each of one or more possible outcomes of transactions, wherein each possible outcome is an end result of processing the transaction;
configuring the classification machine learning model to compute a probability score vector for the transaction, wherein the probability score vector indicates a probability for each of the one or more possible outcomes of the transaction;
analyzing the transaction by the trained causal inference machine learning model, including computing one or more probability mass vectors for the transaction, wherein:
each probability mass vector indicates impact values and associated probabilities of one or more possible treatments to be applied to the transaction, wherein each possible treatment includes a manipulation of the transaction to achieve a favorable outcome of the transaction and includes any one or more of: removing a data field, editing a data field, adding a data field, or reordering the data fields in the transaction;
each impact value indicates a monetary uplift value of applying one possible treatment of the one or more possible treatments; and
each associated probability is a probability of a respective possible monetary uplift value associated with applying the one possible treatment to the transaction;
selecting a treatment to be applied to the transaction by combining an output of the classification machine learning model and an output of the causal inference machine learning model, wherein the selected treatment is based on the probability score vector and the one or more probability mass vectors;
applying the selected treatment to the transaction;
determining a transaction outcome by outputting the transaction after the selected treatment was applied to the transaction; and
feeding back the transaction outcome to the classification machine learning model and the causal inference machine learning model to further train the classification machine learning model and the causal inference machine learning model.
as drafted, constitute a process that, under the broadest reasonable interpretation, covers "certain methods of organizing human activity," specifically, "fundamental economic practices or principles" and/or "commercial or legal interactions," but for recitation of generic computer components. The Examiner notes that "fundamental economic practices" or "fundamental economic principles" describe concepts relating to the economy and commerce, including hedging, insurance, and mitigating risks, and "commercial interactions" or "legal interactions" include agreements in the form of contracts, legal obligations, advertising, marketing or sales activities or behaviors, and business relations. MPEP 2106.04(a)(2)II.A.,B. If a claim limitation, under its broadest reasonable interpretation, covers "fundamental economic practices or principles" and/or "commercial or legal interactions," but for recitation of generic computer components, then it falls within the "certain methods of organizing human activity" grouping of abstract ideas. Accordingly, claims 1, 10 and 17 recite an abstract idea. (Step 2A - Prong 1: YES. The claims recite an abstract idea.)
This judicial exception is not integrated into a practical application. Claims 1, 10 and 17 recite the additional elements of training a classification machine learning model on transaction outcomes of prior transactions; training a causal inference machine learning model on the transaction outcomes of the prior transactions; using the trained classification machine learning model; configuring the classification machine learning model to; by the trained causal inference machine learning model; of the classification machine learning model; of the causal inference machine learning model; and feeding back the transaction outcome to the classification machine learning model and the causal inference machine learning model to further train the classification machine learning model and the causal inference machine learning model (the foregoing recited in claims 1, 10 and 17), and at least one processor and a non-transitory computer-readable medium containing instructions that implement the abstract idea (the foregoing recited in claims 10 and 17), that implement the abstract idea. These additional elements are not described by the applicant and they are recited at a high level of generality (i.e., one or more generic computer elements performing generic computer functions), such that they amount to no more than mere instructions to apply the exception using generic computer elements. Accordingly, even in combination these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. (Step 2A - prong 2: NO. The additional elements do not integrate the abstract idea into a practical application.)
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception itself. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements of training a classification machine learning model on transaction outcomes of prior transactions; training a causal inference machine learning model on the transaction outcomes of the prior transactions; using the trained classification machine learning model; configuring the classification machine learning model to; by the trained causal inference machine learning model; of the classification machine learning model; of the causal inference machine learning model; and feeding back the transaction outcome to the classification machine learning model and the causal inference machine learning model to further train the classification machine learning model and the causal inference machine learning model (the foregoing recited in claims 1, 10 and 17), and at least one processor and a non-transitory computer-readable medium containing instructions that implement the abstract idea (the foregoing recited in claims 10 and 17), to perform the noted steps amount to no more than mere instructions to apply the exception using generic computer elements. Mere instructions to apply an exception using generic computer elements cannot provide an inventive concept ("significantly more"). Accordingly, even in combination, these additional elements do not provide significantly more. As such, claims 1, 10 and 17 are not patent eligible. (Step 2B: NO. The claims do not provide significantly more.)
Dependent claims 2, 3, 5, 7, 11-13, 18 and 21 are similarly rejected because they further define/narrow the abstract idea of independent claims 1, 10 and 17 as discussed above, and/or do not integrate the abstract idea into a practical application or provide an inventive concept such as would render the claims eligible, whether each is considered individually or as an ordered combination.
As for further defining/narrowing the abstract idea:
Claims 2 and 11 merely describe select one or more data fields from the transaction; transform the selected one or more data fields to form a numerical feature vector; and compute the probability score vector based on the numerical feature vector.
Claims 3 and 12 merely describe wherein the treatment parameter is a parameter for determining treatment propensity based on predicted treatment outcomes of the respective plurality of treatments.
Claims 5 and 13 merely describe wherein the … model predicts a marginal likelihood of the user action.
Claim 7 merely describes wherein the one or more … models are trained to predict user actions for one or more of the plurality of treatments based on one or more user features.
Claim 21 merely describes wherein the selecting includes applying a set of decision rules to the probability score vector and the one or more probability mass vectors.
As for additional elements:
Claims 2 and 11 recite "wherein the classification machine learning model is further configured to:" (claims 2 and 11) and “wherein the non-transitory computer-readable medium contains further instructions for the classification machine learning model that, when executed by the at least one processor, cause the at least one processor to:” perform operations (claim 11). This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claim 18 recites “wherein the transaction analysis unit is located at any one or more of: a payment processor, a card network, or an issuing bank.” This recitation generally links the use of a judicial exception to a particular technological environment or field of use. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claims 3, 5, 7, 12, 13 and 21 do not recite any additional elements, and accordingly, for the reasons provided above with respect to the independent claims, are not patent eligible.
Therefore, dependent claims 2, 3, 5, 7, 11-13, 18 and 21 are not patent eligible.
Conclusion
The prior art made of record and not relied upon, as set forth in the accompanying Notice of References Cited (PTO-892), is considered pertinent to applicant's disclosure. Among the cited references:
Yadav (US-20220164798-A1) teaches transaction processing including classifying a transaction using a probability model, based on past transaction outcomes, to predict a current outcome, including generating a probability score vector indicating a probability for each of multiple possible outcomes, selecting and applying a treatment to the transaction based on the analysis/classification, the treatment including manipulating the transaction to achieve a favorable outcome (e.g., enriching the data of the transaction), and determining the current outcome.
Zhao ("Uplift Modeling for Multiple Treatments with Cost Optimization”) teaches uplift modeling, including analyzing a transaction (e.g., a marketing interaction), including computing a probability mass vector indicating impact values and associated probabilities of possible treatments to be applied to the transaction, each treatment including manipulating the transaction to achieve a favorable outcome (e.g., varying a template / changing the communication content of an email), where the impact value indicates a monetary uplift value of applying a possible treatment, and the probability is a probability of the uplift value associated with applying the possible treatment, selecting a treatment to be applied based on the probability mass vector, and applying the selected treatment to the transaction.
Wu (US-20210103926-A1) teaches generating a feature vector from transaction data of a fraudulent transaction and inputting the feature vector into a machine learning model to obtain an output.
Kramme (US-20210065186-A1) teaches classifying a transaction as a certain type of fraud or no fraud, including calculating a score and a probability in that regard.
Cohen (US-20200151825-A1) teaches classifying a transaction into a category such as fraudulent, including generating a transaction vector and a classification vector.
Harris (US-20210234848-A1) teaches using a machine learning model to score a device information vector for probability of fraud.
Zhang (US-20200314101-A1) teaches improving authentication (fraud determination), involving feature vectors, probability determinations, and scoring transactions.
Ozbay (US-20160180228-A1), Gandouet (US-20230196406-A1) and Fahner (US-20110137847-A1) teach uplift modeling in financial contexts.
Rzepakowski ("Decision Trees for uplift modeling with single and multiple treatments"), p. 317 (5 The multiple treatments case), teaches multiple treatments comprising e-mails with different messages advertising the same product, similarly to Zhao ("Uplift Modeling for Multiple Treatments with Cost Optimization").
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DOUGLAS W PINSKY whose telephone number is (571)272-4131. The examiner can normally be reached on 8:30 am – 5:30 pm ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jessica Lemieux can be reached on 571-270-3445. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR.
Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/DOUGLAS W PINSKY/
Examiner, Art Unit 3626
1 Note in Applicant's instant Response, the pagination begins again at 1 on the first page of the Remarks; the page numbers indicated here refer to this re-started pagination of the Remarks section of the Response.
2 Indeed, Applicant's specification (0067-0069) appears to emphasize that the two machine learning models operate separately and independently of each other, and therefore do not interact with each other. (See footnote 4 below.)
3 See footnotes 2 (above) and 4 (below).
4 As noted above, Applicant's specification (0067-0069) appears to emphasize that the two machine learning models operate separately and independently of each other, and therefore do not work "in concert" with each other or interact with each other. Even Applicant's argument here notes this separate/ independent operation of the two machine learning models, which undermines Applicant's assertion that the two machine learning models interact with each other or work in concert/tandem with each other. Rather, the results obtained by the two machine learning models are combined; this interaction occurs in the abstract idea of applying a treatment to the (continued on next page) (continued from previous page) transaction based on analysis of the transaction, not in the additional elements of the two machine learning models.
Prosecution Insights