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 .
DETAILED ACTION
Response to Arguments
Applicant's arguments filed 5/18/2026 have been fully considered, and are partially persuasive. The amendments to independent claims 1, 12, and 19 have – after further search and consideration – resulted in new grounds of rejection presently presented below.
Regarding new claims 22 – 25, Applicant points to paragraphs [46,54,59,63,68] as providing written description support. As discussed in the rejections presented below, these paragraphs - along with consideration of the specification as a whole – fail to provide support for claims 22 - 24.
Specification
The specification is objected to as failing to provide proper antecedent basis for the claimed subject matter for the reasons given below in the 35 USC 112 written description rejection. See 37 CFR 1.75(d)(1) and MPEP § 608.01(o).
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), first paragraph:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 22 – 24 are rejected under 35 U.S.C. 112, first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor(s), at the time the application was filed, had possession of the claimed invention. Regarding claim 22, said claim recites an MLM that “uses the score when the score is calibrated to a loss rate”.
Use of a score “when the score” is of a particular type (i.e., one “calibrated to a loss rate”) requires a consideration/evaluation step to determine if the score is of the particular type (to address the “when” language). Applicant’s specification fails to discuss any consideration/evaluation of that required to support the “uses the score when the score is” language of claim 22. In addition, claim 22 recites when “the score is calibrated to a loss rate”. Discussion of what steps are taken to calibrate a score “to a loss rate” are wholly absent from Applicant’s specification. The use of a “loss rate” is instrumental in establishing a basis for understanding this new claim language. However, “loss rate” only appears a single time in the specification when reciting in [54] “In one embodiment, the model uses scores calibrated to (simulated) loss rates.” How a loss rate is determined and utilized in the manner claimed is not supported, much less how it can be used to determine a score or used in training a machine learning model. Note that even a simple description of what the “loss rate” is directed to is absent; i.e., when understanding a loss rate, one of ordinary skill would first need to understand a loss rate of what, but there is no description as to the source of data used to determine any particular loss rate.
Regarding claim 23, said claim recites “wherein the interpretable MLM uses direct python railyard scoring to produce the score”. “Direct python railyard scoring” is mentioned a single time in [54] of Applicant’s specification, but no description as to how this process is performed or any other description of it is provided. A singular recitation of the phrase “direct python railyard scoring” does not provide written description support for production of a score for use in a MLM in the manner claimed.
Regarding claim 24, said claim recites “wherein the interpretable MLM is further configured to batch the score with one or more other scores to obtain ensembled scores”. “Batch” with context to scores appears a single time in Applicant’s specification in [54]. However, this recitation of a “batch” is prefaced by performing “direct python railyard scoring”, an unsupported operation for the reasons noted above. The “batch” recitation does not appear to be associated with “ensembled scores”; [54] instead discusses “ensembled scores” in association with an operation that “adds together individual leaf values”. “Adds together individual leaf values” does not appear to have any relationship to a “batch” or batching operation given the broadest reasonable interpretation of the “batch”.
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.
Claims 23 and 25 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention. Regarding claim 23, said claim recites use of “direct python railyard scoring”. The specification fails to disclose what “direct python railyard scoring” encompasses. Additionally, “direct python railyard scoring” is not a term of art or a term one of ordinary skill in the art would readily recognize, as evidenced by the lack of results returned from querying the millions of prior art and patent documents in the USPTO’s search tool and the billions of documents indexed by search engines such as Google (both the USPTO SEARCH tool and Google only return results that correspond directly to the present patent application). As a result, the scope of what can be considered “direct python railyard scoring” is both unclear and indefinite.
Regarding claim 25, said claim is recited as further limiting claim 1, particularly further limiting “each binary interaction function”. However, there is no antecedent basis for a “binary interaction function”, and thus it is unclear what aspect of claim 1 is being further specified. Additionally, claim 25 recites that “each binary interaction function of the interpretable MLM is independent”. Claim 25, however, fails to specify what this “independence” is relative to; i.e., the function is “independent” of what exactly? The intended scope of the resultant claim thus is unclear and indefinite.
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 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.
Claims 1, 3, 6, and 7 are rejected under 35 U.S.C. 103 as being unpatentable over Scott (US-11010740-B1) in view of Ramesh (US-20210304204-A1), Tjoa (Tjoa, Erico, et al. "Improving deep neural network classification confidence using heatmap-based eXplainable AI." arXiv preprint arXiv:2201.00009. (Year: 2021)), Manapat (US-10867303-B1), Narsina (Narsina, Deekshith, et al. "AI-driven database systems in fintech: enhancing fraud detection and transaction efficiency." Asian Accounting and Auditing Advancement 10.1: 81-92. (Year: 2019)) and Ekambaram (Ekambaram Vijay, et al. "Attention based multi-modal new product sales time-series forecasting." Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining. (Year: 2020)).
Regarding claim 1, Scott shows a method for one or more processing systems (Fig. 1 item 112 containing systems 120, 122, 124 and 128) reducing use of processing and networking resources for one or more electronic communications (implicit received by achieving the below claimed functionality), the method comprising: receiving, by the one or more processing systems and from a plurality of computing devices, data of network communications performed between the plurality of computing devices (Fig. 1, processing system 112 receives sales, payment, invoice, data from the devices running merchant apps 116 via network 114; see col. 2 lines 9 – 16 and lines 25 – 36, col. 4 lines 10 – 27; this merchant sales data is sourced and communicated via network 114, and used to populate the merchant account information 128) and a plurality of internet-enabled browsers (col. 5 lines 35-45 discussing the merchant device operating an internet (col. 10 lines 56-62) web browser and supports network communication - and utilization of multiple merchant devices - as noted in col. 8 lines 48-65); producing predictions (col. 4 lines 36-66, col. 5 lines 2-25, col. 7 lines 6-27); wherein the plurality of computing devices include one or more computing devices used for one or more communications of communications performed electronically via the Internet (col. 2 lines 49-60, col. 2 line 33 – col. 3 line 24), and
transmitting, based on the prediction (col. 5 lines 2-25, col. 7 lines 6-27), and via a network, a network message to a computing system of the one or more computing devices (a message is sent to merchant 102 from system 112 via network 114 and merchant app 116 as shown in Fig. 1 and discussed via the offer message of col. 3 lines 36 – 58; the offer message can be sent based on the predictions discussed in col. 4 lines 36 – 66; further exchange regarding this offer is illustrated, e.g., in Figs. 3 and 4); receiving, by the one or more processing systems, a network response from the computing system (col. 3 lines 44-52, col. 4 lines 10-15, discussing where the merchant can accept the offer, reject the offer, or request a modified offer (also shown in Fig. 3 step 312); the communication between the merchant 102/110 and the system 112 is via network 114, as illustrated in Fig. 1); and executing, by the one or more processing systems (Fig. 1 item 112, e.g., executing module 122) and based on receiving the network response from the computing system (a merchant response from their device, e.g., Fig. 1 items 102/116), an electronic transfer from the one or more processing systems to a different computer system that maintains an account associated with the computing system (e.g., merchant bank account Fig. 1 item 126, and discussed in col. 3 lines 53-57, col. 9 lines 50-55) by exchanging the one or more electronic communications via a clearing house network (support for clearing house transactions discussed in col. 3 lines 8 – 23; the merchant may be provided a cash advance as discussed in col. 3 lines 53-56; note the repayment rate may be modified based on an updated offer made in response to the predictions noted above, as discussed in col. 7 line 39 – col. 8 line 3, this repayment rate modification would update the electronic payments made to repay the cash advance).
Scott does not show: receiving one or more external signals including an external signal over a network communication channel; training an interpretable machine learning model (MLM) to identify patterns to produce predictions; executing, by the one or more processing systems, the interpretable MLM for the prediction; providing output indicating a behavior of the interpretable MLM in producing the prediction. Ramesh shows: receiving one or more external signals including an external signal over a network communication channel (Fig. 1, [11,17] where agent 110 provides a signal via network 140 regarding, e.g., false positive data); training an interpretable machine learning model (MLM; see Fig. 3B, [12,17,34]) to identify patterns to produce predictions ([15-17] and Fig. 4; the false positive data is utilized along with other transaction data as part of a ML training and retraining process) associated with communications performed electronically via the Internet ([13-16]), executing, by the one or more processing systems, the interpretable MLM for the prediction ([14-16]); providing output indicating a behavior of the interpretable MLM in producing the prediction (Fig. 3B, [17]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the financial processing techniques of Scott with the MLM training, re-training, and evaluation mechanisms of Ramesh in order to improve the utility of the offers extended to the merchant as well as ensure transparency in the underlying processing logic to better ensure the resultant system is operating as intended (Ramesh, [19]).
The above combination does not show providing a heatmap indicating a behavior of the MLM. Tjoa shows use of an interpretable machine learning model (Abstract, pg. 1 L26-L29, L44-L49, Section 2.1, Section A.1), and to provide a heatmap indicating a behavior of the interpretable machine learning model in producing a prediction of the prediction (pg. 1 and pg. 7, L61 – R48). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the interpretable machine learning model of Tjoa in order to ensure the decisions of the MLM can be easily understood (Tjoa, pg. 1 L44-L49), better ensuring this process can be performed reliability and efficiently as the functionality will be better understood by the system’s users. The above combination does not show where the utilized clearing house is an automated clearing house (ACH).
Manapat shows where the utilized clearing house is an automated clearing house (ACH; col. 5 lines 35-45, col. 8 lines 9-21).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the ACH use of Manapat in order to utilize a well-understood prior art environment (the ACH system) for its intended purpose (to efficiently and inexpensively process (bulk) financial transactions), thus better meeting user expectations and ensuring efficient and cost-effective system operations.
The above combination does not show training a machine learning model configured to use a feature based on an external signal of the one or more external signals and produce a prediction that indicates that future processing volume is going to change from processing volume indicated by the data. Narsina shows training a machine learning model (pg. 86 lines 1-18) configured to use a feature (pg. 89 lines 53-63 discussing features such as labels on historical data) based on an external signal of the one or more external signals (pg. 84 lines 20-28 and pg. 86 lines 1-8) and produce a prediction that indicates that future processing volume is going to change from processing volume indicated by the data (pg. 8 line s7-20 and pg. 90 lines 20-24 discussing to “anticipate surges” in view of “historical transaction data”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention with the training techniques of Narsina in order to improve pattern detection capabilities, and thus more reliably identify and predict patterns including fraud and other malicious actions, improving system utility and associated profitability.
The above combination does not show an interpretable model to produce a score and to produce based on the score a prediction. Ekambaram shows an interpretable model to produce a score and to produce based on the score a prediction (pg. 5 Section 3.2.3 and pg. 7, L1 – L27).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention with the additional scoring and forecasting capabilities of Ekambaram that further leverage external data (Abstract, pg. 2 L18-L29 and R33-R40) in order to better anticipate – and as a result provide additional indication – as to whether or not the predicted an actual data will behave as anticipated (Ekambaram, pgs. 6 – 7, Section 4.1).
Regarding claim 3, the above combination further shows wherein the change is a decrease in the future processing volume (Scott, col. 4 line 36 – col. 5 line 25, col. 7 lines 7 – 26 discussing where “lower than expected” volume due to lower inventory levels, lower staffing levels, etc.).
Regarding claim 6, the above combination further shows wherein training the interpretable MLM comprises transforming information associated with transaction data into features that are input into the interpretable MLM (Ramesh, showing various features that are input into the model in Fig. 3A items 1104 and Fig. 3B items 1204, these features include transaction data such as “current amount”, purchases address data (IP and shipping), etc.) and whether the features include the feature (Narsina, pg. 89 lines 53-63).
Regarding claim 7, the above combination further shows wherein the external signal is obtained from a remote location that is remote from the one or more processing systems (Ramesh, Fig. 1 where the Agent Device is remote from Ramesh’s Service Provider Server item 120 and Transaction Processing App 120 and 122 of Ramesh being analogous to the Processing System 112 shown in Scott’s Fig. 1).
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Tjoa, Manapat, Narsina, and Ekambaram, as applied to claim 1 above, further in view of Imrey (US-20110178902-A1). Regarding claim 4, the above combination shows claim 1. The above combination does not show wherein the network communication channel is a File Transfer Protocol (FTP) communication channel. Imrey shows wherein the network communication channel is a File Transfer Protocol (FTP) communication channel ([145]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the FTP utilization of Imrey in order to utilize an old and well-understood data transfer technique (i.e., FTP) for its intended purpose.
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Tjoa, Manapat, Manapat, Narsina, and Ekambaram, as applied to claim 1 above, further in view of Vasudevan (US-20190354895-A1). Regarding claim 5, the above combination shows claim 1, including receiving internal proprietary data associated with the plurality of computing devices and use of the internal proprietary data (Scott, col. 7 lines 7-27 discussing utilizing data such as inventory, payroll, and employee leave) along with use of an interpretable MLM (Ekambaram, pg. 5 Section 3.2.3). The above combination does not show transforming information associated with the data before training the MLM. Vasudevan shows transforming information associated with the data before training the MLM ([6,53-56]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the transformation operations of Vasudevan in order to ensure training data is aligned and the resultant training calibrated based on the characteristics of the provided training data such that the result model’s accuracy is both improved and indicated to those utilizing the model in the future.
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Tjoa, Manapat, Manapat, Narsina, and Ekambaram, as applied to claim 1 above, further in view of Wang (Wang, Chunlan, et al. "Harnessing machine learning emerging technology in financial investment industry: machine learning credit rating model implementation." Journal of Financial Risk Management 10.3: 317-341. (Year: 2021)).
Regarding claim 8, the above combination shows claim 1. The above combination does not show wherein the external signal comprises an option-adjusted spread (OAS) signal.
Wang shows wherein the external signal comprises an option-adjusted spread (OAS) signal (Wang, pg. 7 , lines 1-5).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the external data utilized by Wang in order to leverage well regarded sources of financial information in order to add insights and value to the resultant data production process and ensure its applicability in financial analysis (Wang, pg. 7).
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Tjoa, Manapat, Manapat, Narsina, and Ekambaram, as applied to claim 1 above, further in view of Desai (US-20230206028-A1). Regarding claim 9, the above combination shows setting up the network communication channel to obtain the external signal from a remote location; and receiving the external signal over the network communication channel (Ramesh, showing signaling enabled from 110 to 120/122 via network 140). The above combination does not show reception on a predetermined time interval. Desai shows reception on a predetermined time interval ([14,40] and Figs. 8, 9A suggesting a period, 1-month based process).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the interval-based updating of Desai in order to ensure the evaluated information is current, and thus more likely to produce accurate evaluations.
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Tjoa, Manapat, Manapat, Narsina, and Ekambaram as applied to claim 1 above, further in view of Si (Si, Si, et al. "Gradient boosted decision trees for high dimensional sparse output." International conference on machine learning. PMLR. (Year: 2017)) and Zhu (Zhu, Michael, and Suyog Gupta. "To prune, or not to prune: exploring the efficacy of pruning for model compression." arXiv preprint arXiv:1710.01878. (Year: 2017)). Regarding claim 10, the above combination shows the interpretable machine learning model (Ramesh, Fig. 3A and 3B, and [11-15]; also Tjoa, Abstract, pg. 1 L26-L29, L44-L49, Section 2.1, Section A.1). The above combination does not show use of a constrained model framework implemented with a constrained gradient boosting (GB) tree model (GBTM)
Si shows use of a constrained model framework implemented with a constrained gradient boosting (GB) tree model (GBTM) (pg. 1 and Section 4.2).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the GBTM usage of Si in order to utilize a well-understood ML technique for its intended purpose (i.e., automating data analysis and extrapolation of information from said data).
The above combination does not show utilization of the framework without tree pruning. Zhu shows utilization of the framework without tree pruning (pg. 1).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with Zhu, and weigh pruning versus non-pruning, in order to balance desired accuracy versus model size (i.e., performance).
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Tjoa, Manapat, Manapat, Narsina, and Ekambaram, as applied to claim 1 above, further in view of Boardman (US-20220147817-A1). Regarding claim 11, the above combination shows use of an interpretable MLM (Ramesh, Figs. 3A, 3B, [11-15]). The above combination does not show using monotonicity constraints to ensure an intuitive relationship exists between risk drivers and target. Boardman shows using monotonicity constraints to ensure an intuitive relationship exists between risk drivers and target (Abstract, [1, 5, 16-18, 50]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the monotonic constraints of Boardman in order to, at a low computational cost, improve the ease at which a system user can understand the system’s behavior (Boardman, [16-18]).
Claims 12, 14, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Manapat, Narsina, and Ekambaram.
Regarding claim 12, Scott shows a processing system (Fig. 1, item 112) comprising: a memory to store instructions; and one or more processors, coupled to the memory (Fig. 5), configured to cause the processing system to: receive, from a plurality of computing devices, transaction data of transactions associated with the plurality of computing devices (Fig. 1, processing system 112 receives sales, payment, invoice, data from the devices running merchant apps 116 via network 114; see col. 2 lines 9 – 16 and lines 25 – 36, col. 4 lines 10 – 27; this merchant sales data is sourced and communicated via network 114, and used to populate the merchant account information 128) and a plurality of internet-enabled browsers (col. 5 lines 35-45 and col. 10 lines 56-62 discussing the merchant device operating an internet web browser and support of network communication - and utilization of multiple merchant devices - as noted in col. 8 lines 48-65);
producing predictions (col. 4 lines 36-66, col. 5 lines 2-25, col. 7 lines 6-27); receive, based on the prediction, a network response from a computing system of the plurality of computing devices (col. 3 lines 44-52, col. 4 lines 10-15, discussing where the merchant can accept the offer, reject the offer, or request a modified offer (also shown in Fig. 3 step 312); and execute (Fig. 1 item 112, e.g., executing module 122), based on the network response (a merchant response from their device, e.g., Fig. 1 items 102/116), an electronic transfer from the processing system to a different computer system that maintains an account associated with the computing system (e.g., merchant bank account Fig. 1 item 126, and discussed in col. 3 lines 53-57, col. 9 lines 50-55) by exchanging one or more electronic communications via a clearing house network (support for clearing house transactions discussed in col. 3 lines 8 – 23; the merchant may be provided a cash advance as discussed in col. 3 lines 53-56; note the repayment rate / withheld amount may be modified based on an updated offer made in response to the predictions noted above, as discussed in col. 7 line 39 – col. 8 line 3, this repayment rate modification would update the electronic payments made for the cash advance).
Scott does not show to apply an interpretable machine learning model framework to produce a prediction after the interpretable machine learning model framework is trained using the transaction data. Ramesh shows applying an interpretable machine learning model framework to produce a prediction (Fig. 3B, [12,17,34]) after the interpretable machine learning model framework is trained using the transaction data ([15-17] and Fig. 4; the false positive data is utilized along with other transaction data as part of a ML training and retraining process).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the financial processing techniques of Scott with the MLM training, re-training, and evaluation mechanisms of Ramesh in order to improve the utility of the offers extended to the merchant as well as ensure transparency in the underlying processing logic to better ensure the resultant system is operating as intended (Ramesh, [19]).
The above combination does not show where the utilized clearing house is an automated clearing house (ACH).
Manapat shows where the utilized clearing house is an automated clearing house (ACH; col. 5 lines 35-45, col. 8 lines 9-21).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the ACH use of Manapat in order to utilize a well-understood prior art environment (the ACH system) for its intended purpose (to efficiently and inexpensively process (bulk) financial transactions), thus better meeting user expectations and ensuring efficient and cost-effective system operations.
The above combination does not show training a machine learning model configured to use a feature based on an external signal of the one or more external signals and produce a prediction that indicates that future processing volume is going to change from processing volume indicated by the data. Narsina shows training a machine learning model (pg. 86 lines 1-18) configured to use a feature (pg. 89 lines 53-63 discussing features such as labels on historical data) based on an external signal of the one or more external signals (pg. 84 lines 20-28 and pg. 86 lines 1-8) and produce a prediction that indicates that future processing volume is going to change from processing volume indicated by the data (pg. 8 line s7-20 and pg. 90 lines 20-24 discussing to “anticipate surges” in view of “historical transaction data”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention with the training techniques of Narsina in order to improve pattern detection capabilities, and thus more reliably identify and predict patterns including fraud and other malicious actions, improving system utility and associated profitability.
The above combination does not show an interpretable model to produce a score and to produce based on the score a prediction. Ekambaram shows an interpretable model to produce a score and to produce based on the score a prediction (pg. 5 Section 3.2.3 and pg. 7, L1 – L27).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention with the additional scoring and forecasting capabilities of Ekambaram that further leverage external data (Abstract, pg. 2 L18-L29 and R33-R40) in order to better anticipate – and as a result provide additional indication – as to whether or not the predicted an actual data will behave as anticipated (Ekambaram, pgs. 6 – 7, Section 4.1).
Regarding claim 14, the above combination shows the interpretable machine learning model (Ramesh, Fig. 3B, [17]) is based on internal proprietary data (Scott, col. 7 lines 7-27 discussing utilizing data such as inventory, payroll, and employee leave).
Regarding claim 19, Scott shows or more non-transitory computer readable storage media (Scott, Fig. 5) having instructions stored thereupon which, when executed by a system, cause the system (Fig. 1 item 112 containing systems 120, 122, 124 and 128) to perform operations comprising: receiving, from a plurality of computing devices, data (Fig. 1, processing system 112 receives sales, payment, invoice, data from the devices running merchant apps 116 via network 114; see col. 2 lines 9 – 16 and lines 25 – 36, col. 4 lines 10 – 27; this merchant sales data is sourced and communicated via network 114, and used to populate the merchant account information 128);
produce predictions associated with transactions (col. 4 lines 36-66, col. 5 lines 2-25, col. 7 lines 6-27) including predictions, that future processing volume of transaction is going to change from processing volume indicated by the transaction data (col. 3 lines 36 – 67 and col. 4 lines 38 – 67, note the repayment rate may be modified based on an updated offer made in response to the predictions noted above, as discussed in col. 7 line 39 – col. 8 line 3); and executing (Fig. 1 item 112, e.g., executing module 122) an electronic transfer to a different computer system that maintains an account associated with a computing device, of the one or more computing devices (e.g., merchant bank account Fig. 1 item 126, and discussed in col. 3 lines 53-57, col. 9 lines 50-55), by exchanging one or more electronic communications via a clearing house network (support for clearing house transactions discussed in col. 3 lines 8 – 23; the merchant may be provided a cash advance as discussed in col. 3 lines 53-56; note the repayment rate / withheld amount may be modified based on an updated offer made in response to the predictions noted above, as discussed in col. 7 line 39 – col. 8 line 3, this repayment rate modification would update the electronic payments made for the cash advance). Scott does not show: training an interpretable machine learning model (MLM) to identify patterns to produce predictions;
simulate, after training the interpretable MLM based on the transaction data, the interpretable MLM to produce the prediction.
Ramesh shows: training ([15-17] and Fig. 4; the false positive data is utilized along with other transaction data as part of a ML training and retraining process), an interpretable machine learning model (MLM) to identify patterns to produce predictions;
simulate, after training the interpretable MLM based on the transaction data, the interpretable MLM to produce a prediction ([14-16]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the financial processing techniques of Scott with the MLM training, re-training, and evaluation mechanisms of Ramesh, and thus execute based on Ramesh simulating, in order to improve the utility of the offers extended to the merchant as well as ensure transparency in the underlying processing logic to better ensure the resultant system is operating as intended (Ramesh, [19]).
The above combination does not show where the utilized clearing house is an automated clearing house (ACH).
Manapat shows where the utilized clearing house is an automated clearing house (ACH; col. 5 lines 35-45, col. 8 lines 9-21).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the ACH use of Manapat in order to utilize a well-understood prior art environment (the ACH system) for its intended purpose (to efficiently and inexpensively process (bulk) financial transactions), thus better meeting user expectations and ensuring efficient and cost-effective system operations.
The above combination does not show training a machine learning model configured to use a feature based on an external signal of the one or more external signals and produce a prediction that indicates that future processing volume is going to change from processing volume indicated by the data. Narsina shows training a machine learning model (pg. 86 lines 1-18) configured to use a feature (pg. 89 lines 53-63 discussing features such as labels on historical data) based on an external signal of the one or more external signals (pg. 84 lines 20-28 and pg. 86 lines 1-8) and produce a prediction that indicates that future processing volume is going to change from processing volume indicated by the data (pg. 8 line s7-20 and pg. 90 lines 20-24 discussing to “anticipate surges” in view of “historical transaction data”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention with the training techniques of Narsina in order to improve pattern detection capabilities, and thus more reliably identify and predict patterns including fraud and other malicious actions, improving system utility and associated profitability.
The above combination does not show an interpretable model to produce a score and to produce based on the score a prediction. Ekambaram shows an interpretable model to produce a score and to produce based on the score a prediction (pg. 5 Section 3.2.3 and pg. 7, L1 – L27).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention with the additional scoring and forecasting capabilities of Ekambaram that further leverage external data (Abstract, pg. 2 L18-L29 and R33-R40) in order to better anticipate – and as a result provide additional indication – as to whether or not the predicted an actual data will behave as anticipated (Ekambaram, pgs. 6 – 7, Section 4.1).
Claim 16 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Manapt, Narsina, and Ekambaram, as applied to claim 12 above, further in view of Hennick and Desai.
Regarding claim 16, the above combination does not show wherein the external signal is an option-adjusted spread (OAS) signal. Hennink shows wherein the external signal is an option-adjusted spread (OAS) signal (pg. 215 and pgs. 225-226, suggesting using OAS data from external (i.e., remote) sources such as Merrill Lynch and Moodys for evaluating credit risk). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the datapoints utilized by Hennink in order to make more informed (and thus more accurate) decisions regarding the economic context surrounding the relevant parties involved in the financial transaction and the transaction itself.
The above combination does not show use of a predetermined time interval. Desai shows reception on, and use of, a predetermined time interval ([14,40] and Figs. 8, 9A suggesting a period, 1-month based process).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the interval-based updating of Desai in order to ensure the evaluated information is current, and thus more likely to produce accurate evaluations.
Claim 17 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Manapt, Narsina, and Ekambaram, as applied to claim 12 above, further in view of Si and Zhu.
Regarding claim 17, the above combination shows the above combination shows the interpretable machine learning model (Ramesh, Fig. 3A and 3B, and [11-15]). The above combination does not show use of a constrained model framework implemented with a constrained gradient boosting (GB) tree model (GBTM)
Si shows use of a constrained model framework implemented with a constrained gradient boosting (GB) tree model (GBTM) (pg. 1 and Section 4.2).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the GBTM usage of Si in order to utilize a well-understood ML technique for its intended purpose (i.e., automating data analysis and extrapolation of information from said data).
The above combination does not show utilization of the framework without tree pruning. Zhu shows utilization of the framework without tree pruning (pg. 1).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with Zhu, and weigh pruning versus non-pruning, in order to balance desired accuracy versus model size (i.e., performance).
Claim 21 is rejected under 35 U.S.C. 103 as being unpatentable over Scott in view of Ramesh, Manapt, Narsina, and Ekambaram, as applied to claim 19 above, further in view of Tjoa.
Regarding claim 21, the above combination shows claim 19. The above combination does not show providing provide a heatmap indicating a behavior of the interpretable machine learning model in producing the prediction.
Tjoa shows providing a heatmap indicating a behavior of the interpretable machine learning model in producing the prediction (pg. 1 and pg. 7, L61 – R48). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to modify the above combination with the interpretable machine learning model of Tjoa in order to ensure the decisions of the MLM can be easily understood (Tjoa, pg. 1 L44-L49), better ensuring this process can be performed reliability and efficiently as the functionality will be better understood by the system’s users.
Conclusion
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 JOHN M MACILWINEN whose telephone number is (571)272-9686. The examiner can normally be reached Monday - Friday, 9:00 - 5:00.
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, Glenton B Burgess can be reached at (571) 272 - 3949. 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.
JOHN MACILWINEN
Primary Examiner
Art Unit 2442
/JOHN M MACILWINEN/ Primary Examiner, Art Unit 2454
Prosecution Insights