DETAILED ACTION
1. The present application, filed on or after March 13, 2013, is being examined under the first inventor to file provisions of the AIA .
This is a regular utility application without a claim of priority.
Response to Amendment
2.. An Amendment was filed August 25, 2026 (hereinafter “Amendment”) and has been entered into the record and fully considered. The Amendment was filed in response to a Non- Final Rejection dated June 22, 2026.
However, despite the Amendment to the Claims and Applicant’s remarks, the Rejection under §103 as set forth in the Final Rejection is hereby maintained; although, the rejection is on NEW GROUNDS - necessitated by the Amendment - as explained below.
Eligibility of the Claims under §101 was explained in the Non-Final Rejection.
An explanation of the NEW GROUNDS of Rejection under §103 and a response to Applicant’s arguments are set forth below. Please see the “Conclusion” section of this Action below for important information regarding responding to this Action.
NOTE: Despite numerous suggestions, the Office notes that no interview has been requested or held in this application.
The Examiner believes that, given the eligibility of the Claims under §101, compact prosecution could be better obtained through the interview process. Applicant is encouraged to schedule an interview using the AIR form, the link for which is set forth in the Conclusion section of this Action.
Status of the Claims:
Claims 1, 6 – 11, and 16 – 20 are pending in this Application.
Independent Claim 1 was amended as set forth below and independent Claim 11 is of substantially the same structure and scope and was amended in substantially the same manner as Claim 1.
The dependent Claims were not amended or only amended in a trivial manner.
Therefore, the following explanation of the maintained rejections with regard to Claim 1 is considered explanatory of the Rejection as a whole.
With regard to the Amendment:
Claim 1 was amended as follows:
PNG
media_image1.png
810
706
media_image1.png
Greyscale
PNG
media_image2.png
814
649
media_image2.png
Greyscale
PNG
media_image3.png
578
616
media_image3.png
Greyscale
Summary of the Amendment and Broadest Reasonable Interpretation:
Claim terminology is to be given its plain and ordinary meaning to a person of ordinary skill in the art, consistent with the specification. This is true, unless the terms are given a special meaning. See MPEP §2111.01
Here, no special meaning is detected. As noted in the Amendment, the
changes to Claim 1 relate generally to a multi-stage training process wherein the training samples are discretized into n- bins according to fraud score. This process is common and well-known. It is sometimes referred to as “binning” or discretization or bucketing.
The amended Claim then applies the trained model to detect fraud.
These terms appear – subject to further consideration – to be defined in the specification based on their plain and ordinary meaning.
With regard to §103:
New Grounds of Rejection:
The Claims remain rejected under 35 U.S.C. §103 as being unpatentable over U.S. Patent Publication No. 2022/0379868 to Dherange et al. (hereinafter “Dherange) in view of U.S. Patent Publication No. 2024/0256960 to Chuying et al. (hereinafter “Chuying”) and further in view of U.S. Patent Publication No. 2025/0254185 to Zoldi et al. (hereinafter “Zoldi”) and further in view of Non-Patent Literature Anonymous Blog, “Data discretization in machine learning,” Train In Data, https://blog.trainindata.com 2022 (hereinafter “Train In”)
The following explanation of the maintained rejection is considered explanatory of the rejection as a whole.
As noted above, the changes to Claim 1 essentially involve the combined use of clustering with a binning concept. Scoring is used for each transaction within each bin. The common use of binning – or “discretization” – is then applied to the transaction in each cluster. This process is applied iterative. In the view of the Office, the number of bins utilized and the number of iterations of applying this process are considered arbitrary and obvious. Please see the Non-Final Rejection for further explanation of the common application of binning.
Respectfully, despite Applicant’s Amendment and its good faith attempt to define over the prior art of record, the Office is of the opinion that the added limitations are still clearly taught by the primary reference to Dherange. Out of an abundance of caution, the additional Non-Patent Literature reference to Train In is added to show the common use of clustering in combination with binning.
Dherange teaches the use of clustering the unlabeled transactions:
“[0005] In another example aspect, a computer-implemented method, implemented by at least one computing platforms operating in a cluster-based parallel computing environment; includes acquiring a plurality of records, each record having a corresponding number of attributes, identifying outliers in the plurality of records using labels generated from processing the plurality of records through an ensemble of different deep learning models, wherein an output of at least one model is used as an input to at least one other model; and detecting anomalies in the plurality of records using a probabilistic classifier based on plurality of records and labels.” (Emphasis Added)
It is noted that, rather than the term “features” as used in the Claims, Dherange uses the term “attributes.” This term is considered to constitute the recited term as one of ordinary skill would readily understand.
Dherange teaches that an “ensemble” of ML models is used “wherein the output of at least one model is used as an input to at least one other model.” This is clearly deemed to be a teaching of an iterative approach to labeling previous unlabeled data samples in an unsupervised learning approach.
The cluster-based aspect of Dherange is shown in Fig. 10:
PNG
media_image4.png
462
704
media_image4.png
Greyscale
This Figure and its accompanying description (see at least 0166 – 0170) clearly teach the use of clustering and scoring to identify “anomalies.” The “iterative” nature of this approach is also taught at 0198 – 0199. Binary scoring of 0 – 1 is also taught by this reference at 0171. The use of binning to label the data is shown at 00823 – 0083, as explained in detail in previous actions.
Furthermore, as explained in previous actions, Zoldi teaches that two models are used to determine the high and low scores used in the binning process:
“[0012] An adversary detection model, in accordance with some embodiments, may detect an abnormal volume of activations of at least one adversarial latent feature spanning across a plurality of transactions scored and block transactions that match the adversarial latent feature. The adversary detection model may be implemented based on newly labeled adversarial transactions to determine the latent features describing features relationships in an adversarial space. The one or more extracted features for the first transaction may be binned into ranges of values observed in a training dataset. Various attributes and adversarial latent features of the adversary detection model, including at least one of a neural network or a multi-layered self-calibrated model (MLSC), may be aggregated using moving averages across various transactions that score low by the decision model and high by the adversary detection model in the production system.” (Emphasis Added)
Similarity, Zoldi teaches the iterative application of labeled training data using a binning approach:
Multiple models are used in succession:
“[0020] In some embodiments, the system may comprise a second AI model for identifying adversarial transactions, in response to determining that number of plurality of example transactions scored by the first model is sufficient to train the second AI model, the second AI model being trained using a supervised learning approach based on features associated with the plurality of example transactions scored by the first AI model, the training of the second AI model being according to a stochastic gradient descent method.
[0021] In some embodiments, the system may comprise a third AI model for identifying adversarial transactions, in response to determining the number of the plurality of example transactions scored by the AI model is insufficient to train the second AI model, the MLSC model being trained using a semi-supervised learning approach in which a large corpus of possible latent features associated with adversarial transactions is defined and used to detect adversarial transactions, strength of one or more hidden nodes of the MLSC model being calculated so that the hidden nodes with strength values below a threshold are eliminated to make the MLSC model more efficient.” (Emphasis Added)
Finally, an iterative process is clearly taught by Zoldi:
“[0037] In the initial stages of the learning phase, the categorization may be based on randomly assigned weights and biases, and therefore highly inaccurate. However, learning software 112 may be trained based on certain incentives or disincentives (e.g., a calculated loss function) to adjust the manner in which the provided input is classified. The adjustment may be implemented by way of adjusting weights and biases associated with the input data. Through multiple iterations and adjustments, the internal state of learning software 112 may be continually updated to a point where a satisfactory predictive state is reached (i.e., when learning software 112 starts to more accurately classify the inputted events at or beyond an acceptable threshold).” (Emphasis Added)
The new reference to Train In teaches the common use of clustering in combination with binning. The introduction is as follows
PNG
media_image5.png
722
1038
media_image5.png
Greyscale
Train In identifies the “challenge” in applying discretization:
PNG
media_image6.png
299
1009
media_image6.png
Greyscale
Thus, Train In addresses the same problem as allegedly solved in the claimed invention and teaches the exact same solution:
PNG
media_image7.png
522
1075
media_image7.png
Greyscale
Therefore, it would have been obvious to one of ordinary skill in the relevant art at the time of filing the claimed invention to have modified the combined labeling or tagging system of Dherange in view of Chuying in view of Zoldi to add the combined use of binning and clustering as clearly taught in Train In. The motivation to do so comes from Dherange. As quoted above, Dherange also teaches the use of clustering. It would greatly enhance the efficiency and accuracy of the combined system of Dherange in view of Chuying in view of Zoldi to add the combined use of clustering and binning/discretization.
Therefore the Rejection under §103 must be maintained.
Response to Arguments
3. Relative to §103, Applicant's arguments set forth in the Remarks section of the Amendment are moot in view of the new grounds of Rejection.
The Rejection must be maintained.
Conclusion
4. Applicant should carefully consider the following in connection with this Office Action:
A. Finality
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. 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.
B. Search and Prior Art
The search conducted in connection with this Office Action, as well as any previous Actions, encompassed the inventive concepts as defined in the Applicant’s specification. That is, the search(es) included concepts and features which are defined by the pending claims but also pertinent to significant although unclaimed subject matter. Accordingly, such search(es) were directed to the defined invention as well as the general state of the art, including references which are in the same field of endeavor as the present application as well as related fields (e.g. machine learning to detect fraud). Indeed, there is a plethora of prior art in these fields.
Therefore, in addition to prior art references cited and applied in connection with this and any previous Office Actions, the following prior art is also made of record but not relied upon in the current rejection:
Chinese Patent Publication No. CN 115730254 to Ming et al. This reference relates to the concept of using clustering in combination with binning.
Chinese Patent Publication No. CN 117994049 to Kai. This reference relates to the common concept of using clustering in combination with binning.
C. Responding to this Office Action
In view of the foregoing explanation of the scope of searches conducted in connection with the examination of this application, in preparing any response to this Action, Applicant is encouraged to carefully review the entire disclosures of the above-cited, unapplied references, as well as any previously cited references. It is likely that one or more such references disclose or suggest features which Applicant may seek to claim. Moreover, for the same reasons, Applicant is encouraged to review the entire disclosures of the references applied in the foregoing rejections and not just the sections mentioned.
D. Interviews and Compact Prosecution
The Office strongly encourages interviews as an important aspect of compact prosecution. Statistics and studies have shown that prosecution can be greatly advanced by way of interviews. Indeed, in many instances, during the course of one or more interviews, the Examiner and Applicant may reach an agreement on eligible and allowable subject matter that is supported by the specification.
Interviews are especially welcomed by this examiner at any stage of the prosecution process. Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool (e.g. TEAMS).
To facilitate the scheduling of an interview, the Examiner requests the use of the AIR form as follows:
USPTO Automated Interview Request http://www.uspto.gov/interviewpractice.
Other forms of interview requests filed in this application may result in a delay in scheduling the interview because of the time required to appear on the Examiner's docket. Thus, the use of the AIR form is strongly encouraged.
E. Communicating with the Office
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WILLIAM BUNKER whose telephone number is (571)272-0017. The examiner can normally be reached on M - F 8:30AM - 5:30PM, Pacific.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Abhishek Vyas, can be reached at 571-270-1836. Information regarding the status of an application, whether published or unpublished, may be obtained from the “Patent Center” system. For more information about the Patent Center system, see https://patentcenter.uspto.gov/
/William (Bill) Bunker/
U.S. Patent Examiner
AU 3691
william.bunker@uspto.gov
(571) 272-0017
October 12, 2026
/ABHISHEK VYAS/Supervisory Patent Examiner, Art Unit 3691
