DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, 18/353,743, filed 07/17/2023, claims foreign priority to European Patent Application 23161993.3, filed 03/15/2023.
The effective filing date is after the AIA date of March 16, 2013, and so the application is being examined under the “first inventor to file” provisions of the AIA .
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.
Priority
Acknowledgment is made of applicant's claim for foreign priority, based on European Patent Application 23161993.3, filed 03/15/2023. On 08/09/2023, the certified priority document was electronically retrieved by USPTO from WIPO. The receipt of the certified copies of papers, as required by 37 CFR 1.55, is hereby acknowledged.
Status of the Application
This Final Office Action is in response to Applicant’s communication of 02/09/2025.
Claims 1, 2, and 4-19 are pending, of which claims 1 and 8 are independent.
Claims 1, 7, 8, and 17 have been amended. Claim 3 has been cancelled.
All pending claims have been examined on the merits.
Claim Interpretation
The Examiner has interpreted the expression “associated with” as being synonymous with the word “of”.
The Examiner has interpreted the expression “optionally” (which is recited in claims 7, 9, and 11-13) according to MPEP § 2173.05(h)(II) “Alternative Limitations, Optionally”, which states:
Another alternative format which requires some analysis before concluding whether or not the language is indefinite involves the use of the term "optionally." In Ex parte Cordova, 10 USPQ2d 1949 (Bd. Pat. App. & Inter. 1989) the language "containing A, B, and optionally C" was considered acceptable alternative language because there was no ambiguity as to which alternatives are covered by the claim. A similar holding was reached with regard to the term "optionally" in Ex parte Wu, 10 USPQ2d 2031 (Bd. Pat. App. & Inter. 1989). In the instance where the list of potential alternatives can vary and ambiguity arises, then it is proper to make a rejection under 35 U.S.C. 112(b) and explain why there is confusion.
The Examiner interprets that any prior art reference cited in a 102 or 103 rejection reads upon the “optional” features of the claim, because such features are “optional” (and thus are not obligatory).
The Examiner has interpreted the expression “noised embedding” (which is recited in claims 5 and 11) as defined in the following paragraphs of US 2024/0311834 A1:
[0131] In examples, key and value vectors 1102, 1104 (which, in this example, correspond to the key- value embedding pairs 1012, 1014 described above with reference to FIG. 10 ) are noised 1106 through a differentiable reparameterization procedure, encrypted 1108 and returned to SWIFT as noised embeddings 1110 without identifiers 1112, as shown in FIG. 11.
The Examiner interprets that “key and value embeddings” (which is recited in claim 3) are synonymous with “key and value vectors”, based on the following paragraphs of US 2024/0311834 A1:
[0132] In this example, the same noise 1106 is applied to both the key and value vectors 1102, 1104. In other examples, different noise 1106 may be used.
[0133] In this specific example, key and value vectors 1102, 1104 and an account embedding comprises the key and value vectors 1102, 1104. The key and value vectors 1102, 1104 can be used by an attention mechanism as will be described in more detail below. However, in other examples, an account embedding does not comprise key and value vectors, for example where an attention mechanism is not used.
[0164] The creditor bank node 1510 transmits the noised value and key embeddings 1532, 1534 to the processor node 1506. In this example, the noised value and key embeddings 1532, 1534 are transmitted as an account embedding. An account embedding may also be referred to as an “embedded account representation” or the like.
[0177] The originating bank 1602 uses the sensitive customer data 1608 as input to an originating bank private embedding model 1614. The originating bank 1602 may resample the sensitive customer data 1608 prior to inputting the sensitive customer data 1608 to the originating bank private embedding model 1614. The originating bank private embedding model 1614 outputs key and value embeddings. In examples, the key and value embeddings are based not only on sensitive customer data of the customer that is party to the payment request, but also on sensitive customer data of at least one other customer of the originating bank 1602 that is not a party to the payment request.
The expression “account embedding” (which is recited in claims 1, 3-8, and 11-13) is interpreted in light of the following specific definitions in the specification (i.e. the following paragraphs of US 2024/0311834 A1):
[0118] In terms of bank models, in examples, each bank owns a private model, with a private architecture and/or schema only known to itself. This produces a standardized, pseudo-randomized output, typically referred to herein as an “account embedding”. SWIFT can use the account embedding for the estimation of financial crime prevalence.
[0127] The key embeddings 1012 help SWIFT identify the target account holder. The value embeddings 1014 represent obfuscated de-identified account profiles, with learnt features for anomaly detection.
[0128] In examples, embeddings comply with an agreed dimensionality and encode statistical properties as instructed by SWIFT, or an external orchestrator.
[0129] In examples, a bank is free to choose, privately, the data and model leveraged to produce account embeddings, according to their capabilities and/or preferences.
[0130] FIG. 11 shows an example procedure 1100 to noise and return account embeddings to a payment processor server.
[0131] In examples, key and value vectors 1102, 1104 (which, in this example, correspond to the key-value embedding pairs 1012, 1014 described above with reference to FIG. 10 ) are noised 1106 through a differentiable reparameterization procedure, encrypted 1108 and returned to SWIFT as noised embeddings 1110 without identifiers 1112, as shown in FIG. 11. For example, the communication protocol between SWIFT and the financial institution(s) may attach a unique request ID to each request. This enables requests to be treated as API calls without requiring the hashed and cropped IDs. An embedding that is based on noised data may be referred to as a “noised embedding”.
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, 2, and 4-19 are rejected under 35 U.S.C. §101 because the claimed invention is directed to non-statutory subject matter. The claimed invention is directed to an abstract idea, without “significantly more”.
In regards to Step 1 of the Alice/Mayo analysis, independent claims 1 and 8 are method claims, and dependent claim 15 is an apparatus claim.
For the sake of compact prosecution, we continue with the Alice/Mayo “abstract idea” analysis.
The abstract idea elements recited in independent claim 1 and 8, and dependent claim 15, are shown in italic font. The “additional elements” and “extra solution steps” are shown in underlined and italic font.
In regards to claim 1,
1. (Currently Amended) A method of preserving privacy for a transaction in a federated learning system, the federated learning system comprising a payment decisioning server associated with a payment decisioning entity and a financial institution server associated with a financial institution, the payment decisioning entity and the financial institution being isolated from each other, the transaction involving a customer of the financial institution, the method comprising, at the payment decisioning server:
transmitting, to the financial institution server, a query relating to the customer;
receiving, from the financial institution server, a response to the query, the response comprising an account embedding,
wherein the account embedding is a standardized, pseudo-randomized output generated at the financial institution server by inputting account information associated with the customer into a financial institution private embedding model, the financial institution private embedding model being a differentiable model having an architecture, weights and/or schema private to the financial institution,
wherein the account embedding comprises a key embedding and a value embedding, wherein the key embedding comprises obfuscated customer identifiers, and wherein the value embedding comprises obfuscated de-identified account profiles with learnt features for anomaly detection;
generating a query embedding by inputting transaction data associated with the transaction into a payment decisioning entity private embedding model;
generating a customer embedding by inputting the query embedding, the key embedding and the value embedding into an attention mechanism,
wherein the customer embedding is based on the value embedding and a weighted match of relevant account profiles in the key and query embeddings; and
generating a transaction decision for the transaction by inputting the customer embedding and transaction data associated with the transaction into a payment decisioning entity decision model.
In regards to claim 8,
8. (Currently Amended) A method of preserving privacy for a transaction in a federated learning system, the federated learning system comprising a payment decisioning server associated with a payment decisioning entity and a financial institution server associated with a financial institution, the payment decisioning entity and the financial institution being isolated from each other, the transaction involving a customer of the financial institution, the method comprising, at the financial institution server:
receiving, from the payment decisioning server, a query relating to the customer;
generating, based on the received query, an account embedding by inputting account information associated with the customer as input to a financial institution private embedding model,
wherein the financial institution private embedding model is a differentiable model having an architecture and/or schema private to the financial institution and the account embedding is a standardized, pseudo-randomized output, and wherein generating the account embedding comprises:
generating a key embedding comprising obfuscated customer identifiers; and
generating a value embedding comprising obfuscated de-identified account profiles with learnt features for anomaly detection; and
transmitting, to the payment decisioning server, a response to the received query, the response comprising the account embedding,
wherein the payment decisioning server is configured to:
generate a query embedding by inputting transaction data associated with the transaction into a payment decisioning entity private embedding model;
generate a customer embedding by inputting the query embedding, the key embedding, and the value embedding into an attention mechanism,
wherein the customer embedding is based on the value embedding and a weighted match of relevant account profiles based on the query embedding and the key embedding; and
generate a transaction decision for the transaction by inputting the customer embedding and transaction data associated with the transaction into a payment decisioning entity decision model.
In regards to claim 15,
15. (Original) A server configured to perform a method according to claim 8.
Claims 1, 2, and 4-19 recite “Mathematical Concepts", specifically “Mathematical Relationships”, “Mathematical Formulas or Equations”, and “Mathematical Calculations”, as discussed in MPEP §2106.04(a)(2) Part (IV), and in the 2019 Revised Patent Subject Matter Eligibility Guidance.
The mathematical elements include:
In claim 1: “wherein the account embedding is a standardized, pseudo-randomized output generated at the financial institution server”.
In claim 1: “wherein the account embedding comprises a key embedding and a value embedding, wherein the key embedding comprises obfuscated customer identifiers, and wherein the value embedding comprises obfuscated de-identified account profiles with learnt features for anomaly detection”.
In claim 1: “wherein the customer embedding is based on the value embedding and a weighted match of relevant account profiles in the key and query embeddings”.
In claim 6: “transmitting, to the financial institution server, a loss function gradient with respect to the account embedding”.
In claim 8: “wherein the financial institution private embedding model is a differentiable model having an architecture and/or schema private to the financial institution and the account embedding is a standardized, pseudo-randomized output”.
In claim 8: “wherein the customer embedding is based on the value embedding and a weighted match of relevant account profiles based on the query embedding and the key embedding”.
In claim 12: “deriving a loss function with respect to weights of the financial institution private embedding model based on: the loss function gradient with respect to the account embedding;”.
In claim 12: “a gradient of the account embedding with respect to the weights of the financial institution private embedding model, optionally wherein the method further comprises updating weights of the financial institution private embedding model based on the loss function with respect to weights of the financial institution private embedding model”.
In claim 16: “the identifier associated with the customer is a hashed version of a customer identifier of the customer”.
In claim 17: “the method comprises updating weights of the attention mechanism based on a loss function gradient with respect to weights of the payment decisioning entity private embedding model and weights of a payment decisioning entity decision model”.
In claim 17: “the payment decisioning entity private embedding model and the financial institution private embedding model have different weights, schema and/or architectures from each other”.
In claim 19: “wherein the financial institution private embedding model and the further financial institution private embedding model have different weights, schema and/or architectures from each other”.
The “additional elements” include: “a payment decisioning server” (claims 1 and 8), “a financial institution server” (claims 1 and 8), and “A server configured to perform a method according to claim 8” (claim 15).
The “additional extra-solution elements” include:
In claim 1: “transmitting, to the financial institution server, a query relating to the customer”
In claim 1: “receiving, from the financial institution server, a response to the query, the response comprising an account embedding”.
In claim 1: “inputting account information associated with the customer into a financial institution private embedding model”.
In claim 1: “generating a query embedding by inputting transaction data associated with the transaction into a payment decisioning entity private embedding model”.
In claim 1: “generating a customer embedding by inputting the query embedding, the key embedding and the value embedding into an attention mechanism”.
In claim 1: “generating a transaction decision for the transaction by inputting the customer embedding and transaction data associated with the transaction into a payment decisioning entity decision model”.
In claim 8: “receiving, from the payment decisioning server, a query relating to the customer”.
In claim 8: “generating, based on the received query, an account embedding by inputting account information associated with the customer as input to a financial institution private embedding model”.
In claim 8: “transmitting, to the payment decisioning server, a response to the received query, the response comprising the account embedding”.
In claim 8: “generate a query embedding by inputting transaction data associated with the transaction into a payment decisioning entity private embedding model”.
In claim 8: “generate a customer embedding by inputting the query embedding, the key embedding, and the value embedding into an attention mechanism”.
In claim 8: “generate a transaction decision for the transaction by inputting the customer embedding and transaction data associated with the transaction into a payment decisioning entity decision model”.
In claim 15: “A server configured to perform a method according to claim 8”.
This abstract idea is not integrated into a practical application, because:
The claim is directed to an abstract idea with additional generic computer elements. The generically recited computer elements (“a payment decisioning server”, “a financial institution server”, and “A server configured to perform a method according to claim 8”) do not add a meaningful limitation to the abstract idea, because they amount to simply implementing the abstract idea on a computer. The claim amounts to adding the words "apply it" (or an equivalent) with the abstract idea, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea.
The extra-solution activities (“receiving”, “transmitting”, and “inputting”) do not add a meaningful limitation to the method, as they are insignificant extra-solution activity;
The combination of the abstract idea with the additional elements (generically recited computer elements), and/or with the extra-solution activities, does 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 abstract idea, because:
When considering the elements "alone and in combination" (“a payment decisioning server”, “a financial institution server”, and “A server configured to perform a method according to claim 8”), they do not add significantly more (also known as an "inventive concept") to the exception, because they amount to simply implementing the abstract idea on a computer. Instead, they merely add the words "apply it" (or an equivalent) with the abstract idea, or mere instructions to implement an abstract idea on a computer, or merely use a computer as a tool to perform an abstract idea.
In regards to the extra solution activities (“receiving”,“transmitting”, and “inputting”), these are recognized as such by the court decisions listed in MPEP § 2106.05(d).
More specifically, in regards to the “receiving”,“transmitting”, and “inputting” steps, see the court cases OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network) and (presenting offers and gathering statistics), OIP Techs., 788 F.3d at 1362-63, 115 USPQ2d at 1092-93; buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network).
The Examiner holds that the independent claims “use a computer or other machinery in its ordinary capacity for economic or other tasks (e.g., to receive, store, or transmit data)” or “simply add a general purpose computer or computer components after the fact to an abstract idea”.
All dependent claims are also rejected, because they merely further define the abstract idea.
Response to Amendments
Re: Claim Rejections - 35 USC § 112
The previously presented 35 U.S.C. 112(b) rejections of claims 1, 2, and 4-19 were previously withdrawn, in response to Applicant’s amendments to independent claims 1 and 8.
Re: Claim Rejections - 35 USC § 101
The previously presented 35 U.S.C. 101 rejections are amended, in response to Applicant’s amendments to independent claims 1 and 8.
Re: Claim Rejections - 35 USC § 102
The previously presented 35 U.S.C. 102 rejections were previously withdrawn, in response to Applicant’s amendments to the previously presented claims, and Applicant’s addition of new claims 16-19.
Re: Claim Rejections - 35 USC § 103
The previously presented 35 U.S.C. 103 rejections are withdrawn, in response to Applicant’s most recent amendments to the independent claims 1 and 8.
Conclusion
The art made of record and not relied upon is considered pertinent to applicant's disclosure.
US-2025/0061224-A1 to Poh et al. This reference has an effective filing date of Aug. 15, 2023 and therefore was filed too recently to qualify as prior art. However, it teaches “Electronic protection of sensitive information via data embedding and noise addition” and therefore is relevant to the current application. Also according to the IP.com search it is one of the two most similar references (with a different assignee than the present application).
US-2025/0201267-A1 to Kim et al. This reference has an effective filing date of Feb. 28, 2022 and therefore qualifies as prior art. However, the reference is directed to the non-analogous art of “emotion recognition in real time” in an audio signal. See para. [0062] and [0064]:
[0062] The first multi-modal feature extractor 420 inputs the first feature extracted by the first pre-feature extractor to the 1-D convolutional layer and maps the dimension of the first feature to a preset dimension. The second multi-modal feature extractor 422 inputs the second feature extracted by the second pre-feature extractor to the 1-D convolutional layer and maps the dimension of the second feature to a preset dimension. Here, the converted dimensions of the first and second features may be 40 dimensions, but a specific value are not limited to the present embodiment. The first and second multi-modal feature extractors 420 and 422 may generate the query embedding vector, the key embedding vector, and the value embedding vector by matching dimensions of outputs of the convolutional blocks.
[0064] The first multi-modal feature extractor 420 inputs the query embedding vector, the key embedding vector, and the value embedding vector obtained by multiplying the first embedding vector by the respective weight matrices to the plurality of self-attention layers to extract the first multi-modal feature including temporal correlation information. The second multi-modal feature extractor 422 inputs the query embedding vector, the key embedding vector, and the value embedding vector obtained by multiplying the second embedding vector by the respective weight matrices to the plurality of self-attention layers to extract a second multi-modal feature including temporal correlation information. Here, the number of plurality of self-attention layers included in the first and second multi-modal feature extractors 420 and 422 may be two, but are not limited to the present embodiment.
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 extension fee 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 date of this final action.
Any inquiry concerning this communication or earlier communications should be directed to Examiner Ayal Sharon, whose telephone number is (571) 272-5614, and fax number is (571) 273-1794. The Examiner can normally be reached from Monday to Friday between 9 AM and 6 PM. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Christine M Behncke can be reached on (571) 272-8103. The fax 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.
Sincerely,
/Ayal I. Sharon/
Examiner, Art Unit 3695
May 27, 2026
Prosecution Insights