DETAILED ACTION
This Final Office action is in response to Applicant’s 08/10/2026. Claims 1-20 are pending. The effective filing date of the claimed invention is 03/06/2018.
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 .
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 1, 4, 8, 11, 15, 18 recites the limitation "the at least one fraud case" in line 22. There is insufficient antecedent basis for this limitation in the claim rendering the claim indefinite. Appropriate correction is required. See also claims 4, 11, 18.
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 (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 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Pat. No. 9,954,879 to Sift (“Sift”) in view of U.S. Pat. Pub. No. 2018/0182029 to Vinay (“Vinay”).
With regard to claims 1, 8, and 15, Sift discloses the claimed provider computing system comprising:
a network interface structured to facilitate data communication via a network (see e.g. Sift col. 9, line 17-25);
a database structured to store information associated with accounts held by an institution associated with the provider computing system (see e.g. Sift col. 9 ln 45 – col. 10 ln 16, discussing the storage of event data associated with the digital events; Sift does not explicitly disclose the claimed “database” that stores data with accounts held by institution; See Vinay [0002-4] [0015-19] [0073-78], teaches a financial institution maintaining modeling systems with access to credit-worthiness data, historic transaction data, historic fraudulent data, current transaction data, and transaction accounts used at merchant POS systems. Therefore, it would have been obvious to one of ordinary skill in the fraud assessment art before the effective filing date of the claimed invention to modify Sift’s machine learning fraud mitigation system to operate in Vinay’s financial institution transaction account environment, as “Once credit is extended to card members, their accounts may occasionally be compromised. Financial institutions offering transaction accounts may protect users against fraud by reimbursing account holders for fraudulent charges. As such, financial institutions seek to minimize fraud in order to minimize their losses as a result of reimbursement expenses. Financial institutions may implement risk assessment systems to assess risk, but those systems have traditionally been only as effective as the expert-written rules guiding the systems.” Vinay [0003]); and
a processing circuit comprising a processor and a memory, the processing circuit (see e.g. Sift col. 7 ln 55-65; Sift col 14, ln 25-40) structured to:
receive a plurality of fraud cases, each fraud case associated with transaction data (see e.g. Sift abstract, Sift col. 4 ln 40-58, col 1 ln 13-17 This invention relates generally to the digital fraud and abuse field, and more specifically to a new and useful system and method for generating and implementing digital threat mitigation applications in the digital fraud and abuse field.) and having an initial priority score (see e.g. Sift col 2, ln 50-67);
determine an updated priority score for each fraud case based on the transaction data and case prioritization data (see e.g. Sift col 6, ln 50-65), the case prioritization data comprising a set of rules developed using a machine learning model (see e.g. Sift col 12, ln. 11-15; Sift col. 12 ln 39-50);
assign each fraud case to one of a plurality of queues, each fraud case assembled in a fraud case database (e.g. Sift col 6 ln 48 – col 7 ln 11, where the queue engine can divide the cases up into queues by assigning to different individuals, for instance, see also where “The reviewing queue engine 140 additionally functions to arrange the received triggering digital events according to a priority (e.g., according to highest probability of fraud, based on time of receipt or occurrence, or according to greatest potential loss due to fraud, and the like).” Where these could be considered to be a plurality of queues, breaking up the events into highest probability of fraud, based on time of receipt, and the like.);
perform an action on a first fraud case of the plurality of fraud cases based on a first updated priority score (e.g. Sift col 13-16, and claim 1, score thresholds and evaluation criteria trigger automated disposal actions, including approve, hold, cancel, or routing for further review. Vinay [0016-17] decides whether to approve or decline a transaction request based on the ML risk assessment.) associated with the first fraud case, wherein the action comprises automatically closing the first fraud case and transmitting an alert to a user device of an account holder associated with the first fraud case (e.g. Sift, e.g. published claim 14 col 13, ln 5-25, provides for automatic disposal, including cancellation, and an interactive node may send SMS or push communications to a mobile device of the digital actor associated; Sift does not disclose “closing a fraud case”; Guardian teaches [0229-233] distinguishing between open and closed sessions and supports alerts responsive to risk. The Guardian closed-session architecture corroborates “closing.”), wherein automatically closing the fraud case comprises removing the automatically closed fraud case from the memory (See Guardian at [0568-0572] and Fig. 30, [0572] A separate cleanup thread may remove the records displated in LoginStatsHistory and SessionHistory tables. Therefore, it would have been obvious to one of ordinary skill in the fraud detection art before the effective filing date of modify Sift’s fraud threat mitigation system to include having open sessions and closed session, and the removal of the data from the memory by the cleanup thread, the advantage of the combination being “The movement of selected records is accomplished not as a single task but in batches, which have the benefit of being stopped without jeopardizing a completed batch if RADB becomes busy. The moving thread is controlled by a protocol with RA so that it only loads a batch when RADB is free.” Gaurdian [0572]);
assign a second fraud case of the plurality of the plurality of fraud cases to a fraud agent computing terminal associated with a fraud agent responsive to determining that a second updated priority score associated with the second fraud case is at or above a threshold, wherein assigning the at least one fraud case to the fraud agent computing terminal comprises moving the second fraud case to a cache (e.g. Sift col 6 ln 30—col 7 ln 11, discloses routing/flagging digital events requiring additional scrutiny to a reviewing queue engine when confidence is below a threshold, and the review queue includes manual review by human analysts/experts. Sift does not appear to disclose assigning every event whose priority score exceeds a threshold. Vinay teaches risk thresholds and placing likely positives near the top of an ML-inducing ordering, but no fraud-agent assignment. Gaurdian restricts results to the ‘top risks’ and fetches their facts for display. When combined with Sift’s analyst interface and threshold, the combination supports presenting the selected high-risk case an investigator device. For the assigning comprising moving the second fraud case to a cache, see Gaurdian at [0568-572] where sessiionhistory paid is moved only after the session is closed, enqueues newly closed sessions, moves and dequeues the records, and uses a cleanup thread to remove the corresponding Sessionhistory/loginstateshistory records. See combination above. );
receive an input from the fraud agent computing terminal regarding a disposition of the second fraud case (e.g. Sift abstract, “using digital fraud policy to configure a second computing node comprising a decisioning API or a decisioning computing server to automatically evaluate and automatically select one digital event processing outcome of a plurality of digital event processing outcomes that indicates a disposal of the digital events classified as the digital event type”, col. 9-10); and
restructure the case prioritization data using the machine learning model, wherein the machine learning model is retrained based on the input (e.g. Sift col 7 ln 23-55, The disposal decision generated at the reviewing engine queue 140 together with the review input may, in turn, be converted by the system 100 to useable machine learning input into the machine learning digital fraud detection system 120. Thus, the reviewing queue input and disposal decision may be consumed by the machine learning digital fraud detection system 120 as machine learning training data that may be used to adjust weightings of one or more factors of or add new factors (features) with weightings to the existing machine learning models implemented by the machine learning digital fraud detection system 120 thereby improving the technical capabilities of the machine learning digital fraud detection system 120 to evaluate and determine a digital threat level (e.g., digital threat score) associated with digital event data. Sift col. 9-10, published claims 5-6).
With regard to claims 2, 9, and 16, Sift further discloses the plurality of fraud cases are received from a fraud identification system, and wherein the fraud identification system assigns the initial priority score (see e.g. col. 2 ln 50—col. 3 ln 6, Sift’s score API/machine learning system generates a digital threat score indicating likelihood of fraud or abuse.).
With regard to claims 3, 10, and 17, Sift further discloses where the updated priority score is higher than the initial priority score (Sift, the review queue shown above and disposal decisions may be consumed as training data to adjust weights or add new features to existing machine learning models, improving the systems ability to determine a digital threat level such as a digital threat score. This supports the idea that a later or updated score can differ from an earlier score. Sift does not explicitly state that the updated score is higher than the original score. Sift discloses updating/retraining/and rescoring. When combined with Vinay’s financial institutional aspects, thus in a financial fraud workflow, when later-applied prioritization data or model output identifies a case as more suspicious than initially scored, the updated fraud priority risk score would be higher than the initial score).
With regard to claims 4, 11, and 18, Sift further discloses assigning the at least one fraud case to the fraud agent computing terminal associated with the fraud agent comprises: receiving, from the fraud agent computing terminal, a request to return a highest priority fraud case; determining the highest priority fraud case by identifying the at least one fraud case as having a highest updated priority score; and transmitting the at least one fraud case to the fraud agent computing terminal (see e.g. Sift col 6 ln 48—col 7 ln 11).
With regard to claims 5, 12, and 19, Sift further discloses the processing circuit is further structured to store the plurality of fraud cases and the updated priority score for each fraud case in a central case database (Sift above discussing the queuing ability of such data; Vinay also teaches a financial risk system trained on datasets and applying risk models to financial-service requests to generate risk assessments; see combination above).
With regard to claims 6 and 13, Sift further discloses the machine learning model comprises at least one of a supervised learning model, an unsupervised learning model, or a reinforcement learning model (Sift col 4, ln 40—col 5 ln 10).
With regard to claims 7, 14, and 20, Sift further discloses the machine learning model is retrained based on the input, and wherein the case prioritization data is updated based on an output received from the retrained machine learning model (Sift col 7 ln 10-60).
Response to Arguments
Applicant's arguments filed 08/10/2026 have been fully considered.
Applicant’s arguments with respect to the 103 rejection of claim(s) 1-20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument. See where Guardian is used to cure the deficiencies from the amended language, if needed.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Peter Ludwig whose telephone number is (571)270-5599. The examiner can normally be reached Mon-Fri 9-5.
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, Fahd Obeid can be reached at 571-270-3324. 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.
/PETER LUDWIG/Primary Examiner, Art Unit 3627
Prosecution Insights