DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Drawings
The drawings are objected to because figures 10 and 11 are not clear and blurry to view. Corrected drawing sheets in compliance with 37 CFR 1.121(d) are required in reply to the Office action to avoid abandonment of the application. Any amended replacement drawing sheet should include all of the figures appearing on the immediate prior version of the sheet, even if only one figure is being amended. The figure or figure number of an amended drawing should not be labeled as “amended.” If a drawing figure is to be canceled, the appropriate figure must be removed from the replacement sheet, and where necessary, the remaining figures must be renumbered and appropriate changes made to the brief description of the several views of the drawings for consistency. Additional replacement sheets may be necessary to show the renumbering of the remaining figures. Each drawing sheet submitted after the filing date of an application must be labeled in the top margin as either “Replacement Sheet” or “New Sheet” pursuant to 37 CFR 1.121(d). If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on October 8, 2024, was considered by the examiner. The submission is in compliance with the provisions of 37 CFR 1.97.
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 therefore, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea (mental process) without significantly more.
Claim 1:
Regarding claim 1, in step 1 of the 101-analysis set forth in MPEP 2106, the claim recites
“1. A computer-implemented method comprising: generating, by one or more processors and using an event progression model, an event risk data object for an entity identifier based on an entity feature dataset associated with the entity identifier; receiving, by the one or more processors, a natural language query for the event risk data object; generating, by the one or more processors and using a large language model, a structured data object from the natural language query that comprises a structured data format associated with a defined application programming interface (API) for the event progression model; generating, by the one or more processors and using the event progression model and the structured data object, a simulated event risk data object for the entity identifier based on (i) a modification to the entity feature dataset or (ii) one or more attention values associated with the event risk data object; and initiating, by the one or more processors, the performance of a prediction-based action based on the simulated event risk data object”, and a method is one of the four statutory categories of invention.
In step 2A prong 1 of the 101-analysis set forth in the MPEP 2106, the examiner has determined that the following limitations recite a process that, under the broadest reasonable interpretation, covers a mental process but for recitation of generic computer components:
generating, … an event risk data object for an entity identifier based on an entity feature dataset associated with the entity identifier; (mental process, a person can mentally evaluate and generate an event risk data object from looking at an entity feature dataset associated with the entity identifier , see MPEP 2106.04(a)(2)(III)),
generating, … a simulated event risk data object for the entity identifier based on (i) a modification to the entity feature dataset or (ii) one or more attention values associated with the event risk data object; (mental process, a person can mentally evaluate and generate a simulated event risk data object from changing a part of an entity feature dataset or changing attention values of the event risk data object , see MPEP 2106.04(a)(2)(III)),
generating, … a structured data object from the natural language query that comprises a structured data format associated with a defined application programming interface (API) for the event progression model; (This recites a mental process since a person can mentally create and generate a structured data object with pen and paper, see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
In step 2A prong 2 of the 101-analysis set forth in MPEP 2106, the examiner has determined that the following additional elements do not integrate this judicial exception into a practical application:
A computer-implemented method comprising: … by one or more processors and using an event progression model, (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
receiving, by the one or more processors, a natural language query for the event risk data object; (In step 2A, prong 2, receiving a query recites mere data gathering, which is considered insignificant extra-solution activity – see MPEP 2106.05(g)),
…by the one or more processors and using the event progression model and the structured data object, (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
… by the one or more processors and using a large language model, (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
and initiating, by the one or more processors, the performance of a prediction-based action based on the simulated event risk data object, (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
Since the claim as a whole, looking at the additional elements individually and in combination, does not contain any other additional elements that are indicative of integration into a practical application, the claim is “directed” to an abstract idea.
In step 2B of the 101-analysis set forth in the 2019 PEG, the examiner has determined that the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As discussed above, additional elements iv, vi, vii, and viii recite mere instructions to apply the judicial exception using generic computer components, which are not indicative of significantly more. The additional element v recites mere data gathering, and is considered an insignificant extra-solution activity. In step 2B, this insignificant extra-solution activity is well understood routine and conventional activity, which includes receiving or transmitting data over a network from court case Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016), – see MPEP 2106.05(d) (II)(i)),
Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
Claim 2:
Regarding claim 2, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 2 recites the following additional elements:
2. The computer-implemented method of claim 1, wherein generating the structured data object comprises: inputting the natural language query to a large language model to generate the structured data object, (In step 2A, prong 2, inputting recites mere data gathering, which is considered insignificant extra-solution activity – see MPEP 2106.05(g),). In step 2B, this insignificant extra-solution activity is well understood routine and conventional activity which includes receiving or transmitting data over a network from court case Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016) – see MPEP 2106.05(d) (II)(i),
… wherein the structured data object is generated by translating the natural language query into the structured data object using the large language model, (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)), (In step 2B, this is also considered mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 3:
Regarding claim 3, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 3 recites the following additional element:
3. The computer-implemented method of claim 1, wherein the structured data object comprises a set of computer instructions for interacting with the event progression model, (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)), (In step 2B, this is also considered mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 4:
Regarding claim 4, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 4 recites the following additional element:
4. The computer-implemented method of claim 1, wherein generating the structured data object comprises: generating the structured data object by executing a read-evaluate-print (REPL) loop based on the natural language query, (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)), (In step 2B, this is also considered mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 5:
Regarding claim 5, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 5 recites the following abstract ideas:
5. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: removing one or more features from the entity feature dataset based on the structured data object, (This recites a mental process, since a person can mentally evaluate and remove variables or features from a dataset using pen and paper, see MPEP 2106.04(a)(2)(III)),
and generating an event risk score for the entity identifier based on a sequence of removal of the one or more features from the entity feature dataset, (This recites a mental process, a person can mentally evaluate and generate an event risk score from removal of one or more features from a dataset, see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 6:
Regarding claim 6, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 6 recites the following abstract idea:
6. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: modifying a particular feature from the entity feature dataset in response to a determination, based on the structured data object, that a modification to the particular feature results in a greatest change to the event risk data object as compared to one or more other features from the entity feature dataset, (This recites a mental process, a person can mentally evaluate and generate a simulated event risk data object from modifying one or more features from a dataset that gives the greatest change to the event risk data object, see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 7:
Regarding claim 7, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 7 recites the following abstract idea:
7. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: generating a simulated entity feature dataset by adding a simulated event feature to the entity feature dataset based on the structured data object (This recites a mental process, a person can mentally evaluate and generate a simulated dataset with pen and paper by adding a simulated event feature to the entity feature dataset from a structured data object, see MPEP 2106.04(a)(2)(III)),
and generating an event risk score for the entity identifier based on the simulated entity feature dataset, (This recites a mental process, a person can mentally evaluate and generate an event risk score (which is a numeric value) for an entity identifier based on the simulated entity feature dataset, see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 8:
Regarding claim 8, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 8 recites the following abstract ideas:
8. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: determining importance of respective features of the entity feature dataset based on the structured data object (This recites a mental process, a person can mentally evaluate and generate a simulated event risk data object by determining importance of features in a dataset by looking at the structured data object, see MPEP 2106.04(a)(2)(III)),
generating the one or more attention values based on the importance of respective features; (This recites a mental process, a person can mentally evaluate and generate attention values based on importance of respective features, see MPEP 2106.04(a)(2)(III)),
generating a weighted entity feature dataset by weighting the entity feature dataset based on the one or more attention values; (This recites a mental process, a person can mentally evaluate and generate a weighted entity feature dataset from attention values (which are viewed as weights or numeric values from specification paragraph [0086] stating “The one or more attention values may respectively assign a weight to one or more features of the event risk data object 406 based on determined relevance to the natural language query 504,” see MPEP 2106.04(a)(2)(III)),
and generating, … the simulated event risk data object based on the weighted entity feature dataset. (This recites a mental process, a person can mentally evaluate and generate a simulated event risk data object with pen and paper based on the weighted entity feature dataset , see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea. Further, claim 8 recites the following additional elements:
…using the event progression model, (In step 2A, prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)), (In step 2B, this is also considered mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 9:
Regarding claim 9, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 9 recites the following abstract idea:
9. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: generating an event risk score for respective features of the entity feature dataset based on the structured data object (This recites a mental process, a person can mentally evaluate and generate an event risk score for each features of an entity feature dataset, see MPEP 2106.04(a)(2)(III)),
and generating the simulated event risk data object based on the event risk score. (This recites a mental process, a person can mentally evaluate and generate a simulated event risk data object with pen and paper from looking at an event risk score, see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 10:
Regarding claim 10, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 10 recites the following abstract idea:
10. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: generating a modified entity feature dataset by modifying one or more features of the entity feature dataset based on the structured data object (This recites a mental process, a person can mentally evaluate and modify the dataset to generate a modified entity feature dataset by using pen and paper, see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Further, claim 10 recites the following additional element:
and inputting the modified entity feature dataset to the event progression model to generate the simulated risk data object. (In step 2A, prong 2, this recites mere data inputting, which is considered insignificant extra-solution activity – see MPEP 2106.05(g),). In step 2B, this insignificant extra-solution activity is well understood routine and conventional activity which includes receiving or transmitting data over a network from court case Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016) – see MPEP 2106.05(d) (II)(i),
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 11:
Regarding claim 11, in step 1 of the 101-analysis set forth in MPEP 2106, the claim recites
“11. A computing system comprising memory and one or more processors communicatively coupled to the memory, the one or more processors configured to: generate, using an event progression model, an event risk data object for an entity identifier based on an entity feature dataset associated with the entity identifier …”, and a system is one of the four statutory categories of invention.
In step 2A prong 1 of the 101-analysis set forth in the MPEP 2106, the examiner has determined that the following limitations recite a process that, under the broadest reasonable interpretation, covers a mental process but for recitation of generic computer components:
generate, … an event risk data object for an entity identifier based on an entity feature dataset associated with the entity identifier; (mental process, a person can mentally evaluate and generate an event risk data object from looking at an entity feature dataset associated with the entity identifier , see MPEP 2106.04(a)(2)(III)),
generate, … a simulated event risk data object for the entity identifier based on (i) a modification to the entity feature dataset or (ii) one or more attention values associated with the event risk data object; (mental process, a person can mentally evaluate and generate a simulated event risk data object from changing a part of an entity feature dataset or changing attention values of the event risk data object , see MPEP 2106.04(a)(2)(III)),
generate, … a structured data object from the natural language query that comprises a structured data format associated with a defined application programming interface (API) for the event progression model; (This recites a mental process since a person can mentally create and generate a structured data object with pen and paper, see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
In step 2A prong 2 of the 101-analysis set forth in MPEP 2106, the examiner has determined that the following additional elements do not integrate this judicial exception into a practical application:
A computing system comprising memory and one or more processors communicatively coupled to the memory, the one or more processors configured to:… (In step 2A, prong 2, this is considered a generic computer component being used as a tool. – see MPEP 2106.05(f)),
receive a natural language query for the event risk data object; (In step 2A, prong 2, receiving a query recites mere data gathering, which is considered insignificant extra-solution activity – see MPEP 2106.05(g)),
… using a large language model, … (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
… using the event progression model and the structured data object, (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
and initiate the performance of a prediction-based action based on the simulated event risk data object, (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
Since the claim as a whole, looking at the additional elements individually and in combination, does not contain any other additional elements that are indicative of integration into a practical application, the claim is “directed” to an abstract idea.
In step 2B of the 101-analysis set forth in the 2019 PEG, the examiner has determined that the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As discussed above, additional elements vi, vii, and viii recite mere instructions to apply the judicial exception using generic computer components, and additional element iv recites a generic computer component being used as a tool, which are not indicative of significantly more. The additional element v recites mere data gathering, and is considered insignificant extra-solution activity. In step 2B, this insignificant extra-solution activity is well understood routine and conventional activity, which includes receiving or transmitting data over a network from court case Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016), – see MPEP 2106.05(d) (II)(i)),
Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
Claims 12-19:
Since claims 12-19 recite similar limitations as corresponding claims 2-9, respectively listed above, they are rejected for similar reasons under 35 U.S.C. 101.
Claim 20:
Regarding claim 20, in step 1 of the 101-analysis set forth in MPEP 2106, the claim recites
“One or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors, cause the one or more processors to: generate, using an event progression model, an event risk data object for an entity identifier based on an entity feature dataset associated with the entity identifier; receive a natural language query for the event risk data object; generate, using a large language model, a structured data object from the natural language query that comprises a structured data format associated with a defined application programming interface (API) for the event progression model; …”, and a non-transitory computer-readable storage media recites a system and is considered to be one of the four statutory categories of invention.
In step 2A prong 1 of the 101-analysis set forth in the MPEP 2106, the examiner has determined that the following limitations recite a process that, under the broadest reasonable interpretation, covers a mental process but for recitation of generic computer components:
generate, … an event risk data object for an entity identifier based on an entity feature dataset associated with the entity identifier; (mental process, a person can mentally evaluate and generate an event risk data object from looking at an entity feature dataset associated with the entity identifier , see MPEP 2106.04(a)(2)(III)),
generate, … a simulated event risk data object for the entity identifier based on (i) a modification to the entity feature dataset or (ii) one or more attention values associated with the event risk data object; (mental process, a person can mentally evaluate and generate a simulated event risk data object from changing a part of an entity feature dataset or changing attention values of the event risk data object , see MPEP 2106.04(a)(2)(III)),
generate, … a structured data object from the natural language query that comprises a structured data format associated with a defined application programming interface (API) for the event progression model; (This recites a mental process since a person can mentally create and generate a structured data object with pen and paper, see MPEP 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
In step 2A prong 2 of the 101-analysis set forth in MPEP 2106, the examiner has determined that the following additional elements do not integrate this judicial exception into a practical application:
20. One or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors, cause the one or more processors to:… (In step 2A, prong 2, this is considered a generic computer component being used as a tool. – see MPEP 2106.05(f)),
receive a natural language query for the event risk data object; (In step 2A, prong 2, receiving a query recites mere data gathering, which is considered insignificant extra-solution activity – see MPEP 2106.05(g)),
… using a large language model, … (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
… using the event progression model and the structured data object, (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
and initiate the performance of a prediction-based action based on the simulated event risk data object, (In step 2A prong 2, this recites mere instructions to apply an exception using generic computer – see MPEP 2106.05(f)),
Since the claim as a whole, looking at the additional elements individually and in combination, does not contain any other additional elements that are indicative of integration into a practical application, the claim is “directed” to an abstract idea.
In step 2B of the 101-analysis set forth in the 2019 PEG, the examiner has determined that the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As discussed above, additional elements vi, vii, and viii recite mere instructions to apply the judicial exception using generic computer components, and additional element iv recites a generic computer component being used as a tool, which are not indicative of significantly more. The additional element v recites mere data gathering, and is considered insignificant extra-solution activity. In step 2B, this insignificant extra-solution activity is well understood routine and conventional activity, which includes receiving or transmitting data over a network from court case Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016), – see MPEP 2106.05(d) (II)(i)),
Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
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, 2, 11, 12, and 20 are rejected as being unpatentable under 35 U.S.C. 103 over Gnanasambandam, N. et al., in US PG Pub. No. US20230115939A1, published on April 13, 2023, (hereafter, Gnanasambandam), in view of Patil S., et al., in “Gorilla: Large language model connected with massive apis,” published on May 24, 2023, available at: https://arxiv.org/pdf/2305.15334, (hereafter, Patil).
Claim 1:
Regarding claim 1, Gnanasambandam teaches “1. A computer-implemented method comprising: generating, by one or more processors and using an event progression model, an event risk data object for an entity identifier based on an entity feature dataset associated with the entity identifier;”
See Gnanasambandam in [0123] describe “Accordingly, some embodiments provide a technical solution to tracking, monitoring, and proactively performing preventative actions to reduce the impact of infectious diseases. In some embodiments, a cognitive intelligence platform may use an artificial intelligence engine that constantly updates its machine learning models with new knowledge obtained about the infectious disease… Weights may be assigned to various risk factors based on whether the authoritative sources agree on their level of risk.” Here, Gnanasambandam describes using a model (i.e. event progression model that monitors impact of infectious disease risk), where the level of risk relates to an event risk data object.
Further, see Gnanasambandam in abstract mention “A method for monitoring a risk of infectious disease for a patient using a patient viewer provided by a cognitive intelligence platform is disclosed. The method includes receiving a selection to estimate the risk of infectious disease for the patient, presenting, via the patient viewer, a question pertaining to a symptom of the infectious disease in a conversational interface, receiving an answer to the question, transmitting the answer, receiving, from the cognitive intelligence platform, the risk for the infectious disease, where the risk is determined in real-time by the cognitive intelligence platform based on a combination of the answer and a patient graph of the patient maintained by the cognitive intelligence platform, the patient graph including another symptom of the patient obtained from an electronic medical record system” Here, Gnanasambandam describes a method to monitor risk of developing a form of infectious disease for a patient (i.e. entity identifier) based on an electronic medical record system of patients (i.e. relates to an entity feature dataset for the entity). Examiner construes event risk data object to be any data or record information about a predicted risk. See specification in [0062] for details “in some embodiments, the term “event risk data object” refers to a data entity that describes a machine learning prediction for event risk for respective predefined events. In some embodiments, an event risk data object may indicate a predicted degree of risk for respective events. In some embodiments, an event risk data object may include one or more event risk scores for one or more potential events. An event risk score may provide a predicted risk or probability that a particular event will occur at a future instance in time.” Here, the data object is about the risk determined by the cognitive intelligence platform.
Further, Gnanasambandam describes in [0123] “Accordingly, some embodiments provide a technical solution to tracking, monitoring, and proactively performing preventative actions to reduce the impact of infectious diseases. In some embodiments, a cognitive intelligence platform may use an artificial intelligence engine that constantly updates its machine learning models with new knowledge obtained about the infectious disease… Weights may be assigned to various risk factors based on whether the authoritative sources agree on their level of risk.” Here, Gnanasambandam describes using a model (i.e. event progression model that monitors impact of infectious disease risk),
Also, see Gnanasambandam in paragraph [0558] mentions “FIG. 65 shows an example of providing a user interface 6500 depicting the factors used when calculating a risk for a medical condition, in accordance with various embodiments. In the user interface, the risk score for a severe community-acquired pneumonia (SCAP) is being determined.” Here, Gnanasambandam shows a risk score value associated with the interface. See Gnanasambandam in [0559] for details.
Further, see Gnanasambandam in [0004] describe “In some embodiments, a system includes a memory storing instructions and a processor communicatively coupled with the memory. The processor may execute the instructions to perform one or more of the operations described above.” Here, Gnanasambandam shows a processor that helps run the system.
Further, Gnanasambandam teaches “receiving, by the one or more processors, a natural language query for the event risk data object;”
See Gnanasambandam in [0290-0291] describe “FIG. 15 shows a method (1500), in accordance with various embodiments, for answering a user-generated natural language medical information query based on a diagnostic conversational template. [0291] In the method as shown in FIG. 15 , an artificial intelligence-based diagnostic conversation agent receives a user-generated natural language medical information query as entered by a user through a user interface on a computer device (FIG. 15 , block 1502). In some embodiments, the artificial intelligence-based diagnostic conversation agent is the conversation agent 110 of FIG. 1 . In some embodiments the computer device is the mobile device 104 of FIG. 1 . One example of a user-generated natural language medical information query as entered by a user through a user interface is the question “Is a blood sugar of 90 normal?” as shown in line 402 of FIG. 4”. Here, Gnanasambandam shows a user receiving a natural language query about medical information where user ask questions about a medical topic such as is a certain amount of blood sugar level normal (i.e. relates to being part of event risk data object since a data object contains information about one patient or one medical record for a person ). See Gnanasambandam in [0149] for details.
Also, see Gnanasambandam in [0301] mention “In some embodiments, selecting a diagnostic fact variable set relevant to generating a medical advice query answer for the user-generated natural language medical information query by classifying the user-generated natural language medical information query into one of a set of domain-directed medical query classifications associated with respective diagnostic fact variable sets (FIG. 15 , block 1504).” Here, Gnanasambandam shows using the system to respond to the user query.
Also, see Gnanasambandam in [0128] describe “Providers may act more effectively and efficiently based on the collection of risk estimates that are provided in real-time clinical monitoring and status displays. For example, a user interface may be provided in a clinic viewer that presents aggregate risk information for a multitude of patients as well as a geographical map depicting where the patients are located and their respective risk levels. The user interface also includes various preventative actions the medical personnel may perform in real-time to proactively attempt to prevent the person from contracting the infectious disease. Providing the risk information, distribution of people in the geographic map, as well as the preventative actions, an enhanced user interface is provided that may enhance the user experience using the computing device... Accordingly, computing resources may be saved by not performing additional queries or switching between user interface screens, since the relevant information is presented in the single user interface on the clinic viewer.” Here, Gnanasambandam shows that the user can interact with the interface to receive a query to view the risk measures of various patients.
Further, see Gnanasambandam in [0004] describe “In some embodiments, a system includes a memory storing instructions and a processor communicatively coupled with the memory. The processor may execute the instructions to perform one or more of the operations described above.” Here, Gnanasambandam shows a processor that helps run the system. Further, see Gnanasambandam describe in [0538] for details.
Further, Gnanasambandam teaches “generating, by the one or more processors and using the event progression model …, a simulated event risk data object for the entity identifier based on (i) a modification to the entity feature dataset or (ii) one or more attention values associated with the event risk data object;”
See Gnanasambandam in [0545 -0546] describe “The risk may be determined based on factors indicative of being infected by the infectious disease. In some embodiments, medical personnel (medical doctors) may provide their opinion or feedback as to what factors are risk factors for contracting the infectious disease. The medical personnel may indicate some risk factors are more serious than others. The cognitive intelligence platform may update weights on the risk factors in a knowledge graph representing the infectious disease based on a consensus of medical personnel agreeing on which risk factors are serious and which risk factors are not. The opinions of the doctors may change as new information is learned about the infectious disease. Accordingly, the weights may dynamically change such that the risk determination for the patient dynamically changes as the cognitive intelligence platform continuously learns from the medical personnel. [0546] At block 6212, the processing device may present the risk for the infectious disease in the conversational interface of the patient viewer on the computing device of the patient.” Here, Gnanasambandam shows that weights or attention values may be modified depending on new information given by medical personnel, where new information relates to event risk data object. From the specification paragraph [0091] stated, “The one or more attention values may be one or more weights for one or more features of the event risk data object 406,” where the specification shows weights to be synonymous with attention values. Further, Gnanasambandam shows that the medical record of a patient represents the entity feature dataset.
See Gnanasambandam in paragraph [0108] describe “the autonomous multipurpose application may also manage and store other information for the users. For example, the user may capture an image of their driver’s license, insurance card, and the like, and transmit the image to the autonomous multipurpose application... The autonomous multipurpose application may electronically fill information in corresponding documents based on the extracted information. Further, the autonomous multipurpose application may perform logic based on the extracted information. For example, if the user’s insurance is about to expire, the autonomous multipurpose application may transmit a message (e.g., email, text message, phone call, onscreen notification, etc.) to the user to renew their insurance. Similar types of information may be managed and stored for each person in a family. The information may be disbursed to a requesting client, such as an EMR system used by an entity at which the users make appointments.” Here, Gnanasambandam illustrates the EMR system as part of the entity feature dataset that helps organize information related to patients.
Further, see Gnanasambandam in paragraph [0531] mention “at block 5902, the processing device may receive information corresponding to a health artifact of the set of health artifacts in the first data structure... In some embodiments the information may be received from a source including an electronic medical records system, an application programming interface, ...” Here, Gnanasambandam mentions that the data structure or dataset contain information from a medical records which relate to a modification to the entity feature dataset, since receiving information is part of a modification to the dataset of medical records of patient information.
Also, see Gnanasambandam in paragraph [0479] mention “the processing device may receive the portion of the required information and update the one of the respective subsets of the set of check-in documents with the portion of the required information. Further, the processing device may check-in the user for the one of the set of schedule appointments once the update is complete.” Here, Gnanasambandam mentions updating information of a subset (i.e. part of entity feature dataset). Entity feature dataset is construed to mean any data or records that collects information on various variables or features, such as medical records for patients.
Further, Gnanasambandam teaches “and initiating, by the one or more processors, the performance of a prediction-based action based on the simulated event risk data object.”
See Gnanasambandam in [0547] describe “At block 6214, responsive to the risk for the infectious disease satisfying a threshold level (e.g., medium, high, critical), the processing device may perform a preventative action. The preventative action may be selected in proportion to which threshold level is satisfied. A more severe preventative action (e.g., contact emergency services, transmit test kit, etc.) may be performed if the risk level satisfies a high or critical threshold level.” Here, Gnanasambandam shows that depending on the level of risk of the infectious disease, a preventative action may be performed based on predicted risk level.
Also, see Gnanasambandam in [0545-0546] describe “the risk may be determined based on factors indicative of being infected by the infectious disease. In some embodiments, medical personnel (medical doctors) may provide their opinion or feedback as to what factors are risk factors for contracting the infectious disease … The opinions of the doctors may change as new information is learned about the infectious disease. Accordingly, the weights may dynamically change such that the risk determination for the patient dynamically changes as the cognitive intelligence platform continuously learns from the medical personnel.
[0546] At block 6212, the processing device may present the risk for the infectious disease in the conversational interface of the patient viewer on the computing device of the patient.” The examiner construes simulated event risk data object to mean any predicted score or measure either created from a model or other method. Here, Gnanasambandam shows that risk is predicted by the intelligence platform system and present this as simulated or potential risk score for the patient.
However, Gnanasambandam did not teach “generating, by the one or more processors and using a large language model, a structured data object from the natural language query that comprises a structured data format associated with a defined application programming interface (API) for the event progression model;” or “… and using the event progression model and the structured data object, …”
In an analogous art, Patil teaches “generating, by the one or more processors and using a large language model, a structured data object from the natural language query that comprises a structured data format associated with a defined application programming interface (API) for the event progression model;”
See Patil in page 6, section 4. Evaluation, subsection Retrievers mention “The sole input to the model is the user’s natural language prompt. For BM25, we consider each API as a separate document. During retrieval, we use the user’s query to search the index and fetch the most relevant (top-1) API." Here, Patil shows each user is creating an NLP query which gets input into the API..
Also, see Patil in page 1, abstract, mention “We release Gorilla, a fine[-]tuned LLaMA-based model that surpasses the performance of GPT-4 on writing API calls. When combined with a document retriever, Gorilla demonstrates a strong capability to adapt to test-time document changes, enabling flexible user updates or version changes.” Here, Patil shows that the language model Gorilla, accounts for any updates, changes, and works as an event progression model that talks with an API interface. Examiner construes event progression model to be any model that can predict outcomes.
Further, see Patil in page 12, section 8 Appendix, 8.1 Dataset details describe “To enhance the value and utility of our dataset, we’ve undertaken an additional initiative. With each API, we have generated a set of 10 unique instructions. These instructions, carefully crafted and meticulously tailored, serve as a guide for both training and evaluation. This initiative ensures that every API is not just represented in our dataset, but is also comprehensively understood and effectively utilizable. In essence, our dataset is more than just a collection of APIs across three domains. It is a comprehensive resource, carefully structured and enriched with added layers of guidance and evaluation parameters.” Here, Patil describes generating unique instructions for each API as well as generating datasets, which relate to generating a structured data object that is associated with a defined API for a model.
Further, see Patil in page 4, first paragraph mention “we then converted the model cards for each of these 1,645 API calls into a json object with the following fields: {domain, framework, functionality, api_name, api_call, api_arguments, environment_requirements, example_code, performance, and description.}. We provide more information in the Appendix. These fields were chose to generalize beyond the API calls within ML domain, to other domains, [including] RESTful API calls.” Here, Patil shows that the json object is the structured data object, which is associated with the API call. Examiner construes structured data object to be any format that contains data or information.
Further, Patil teaches “… and using the event progression model and the structured data object, …”
See Patil in page 1, abstract, mention “We release Gorilla, a fine[-]tuned LLaMA-based model that surpasses the performance of GPT-4 on writing API calls. When combined with a document retriever, Gorilla demonstrates a strong capability to adapt to test-time document changes, enabling flexible user updates or version changes.” Here, Patil shows that the model Gorilla, accounts for any updates, changes, and works as an event progression model that talks with an API interface.
Further, see Patil in page 4, first paragraph mention “we then converted the model cards for each of these 1,645 API calls into a json object with the following fields: {domain, framework, functionality, api_name, api_call, api_arguments, environment_requirements, example_code, performance, and description.}. We provide more information in the Appendix. These fields were chose to generalize beyond the API calls within ML domain, to other domains, [including] RESTful API calls.” Here, Patil shows that the json object is the structured data object, which is associated with the API call. Examiner construes structured data object to be any format that contains data or information, where in this case is a json object.
Also, see Patil from page 1, abstract discuss about event progression model, and in page 12, section 8, mention about the structured data object.
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Gnanasambandam and incorporate into the teachings of Patil because both references teach using an event progression model to process a natural language query to create a simulated event risk data object.
One of ordinary skill in the art would be motivated to do so because performing “Retriever-Aware training For training with retriever, the instruction-tuned dataset, also has an additional "Use this API documentation for reference: " appended to the user prompt. Through this, we aim to teach the LLM to parse the second half of the question to answer the first half. We demonstrate that this a) makes the LLM adapt to test-time changes in API documentation, and b) improves performance from in-context learning, and finally c) show that it reduces hallucination error,” (see Patil in page 5, paragraph 2).
Claim 2:
Regarding claim 2, Further, Patil teaches “2. The computer-implemented method of claim 1, wherein generating the structured data object comprises: inputting the natural language query to a large language model to generate the structured data object, wherein the structured data object is generated by translating the natural language query into the structured data object using the large language model.”
See Patil in page 6, section 4. Evaluation, subsection Retrievers mention " The sole input to the model is the user’s natural language prompt. For BM25, we consider each API as a separate document. During retrieval, we use the user’s query to search the
index and fetch the most relevant (top-1) API. " Here, Patil explicitly shows each user is creating an NLP input, where the input contains a prompt and user query, which are the inputs into the API interface. The query here is a specific question or input the interface gathered from the user. The prompt is the information sent to the system. When the user input is received by the interface here, Patil shows processing the user input, from both query and prompt, into searching the index and then fetching the most relevant API.
Further, see Patil in page 2, figure 1 mention " our Gorilla model can identify the task correctly and suggest a fully-qualified API call." Here, Patil shows to translate the task (which includes the query) into a format that the API can understand (where API call relates to structured data object). Here, Patil is retrieving the API document as the structured data object in response to the user input by query using the Gorilla model (i.e. natural language model). Here, Patil explicitly mentions using a user query then translate that into an API call to retrieve the relevant information in response to that user input. Here, the term ‘user’s query’ is interpreted as the natural language query. The term ‘API call’ is construed to be a structured data object here.
Further, see Patil in page 2, second paragraph describe “We first parse the generated code into an AST tree, then find a sub-tree whose root node is the API call that we care about (e.g., torch.hub.load) and use it to index our dataset. We check the functional correctness and hallucination problem for the LLMs, reporting the corresponding accuracy.” Here, Patil shows the process of how to translate the natural language query or prompt into a language that the API understands using Gorilla model (i.e. large language model).
Also, see Patil in page 4, first paragraph mention “we then converted the model cards for each of these 1,645 API calls into a json object with the following fields: {domain, framework, functionality, api_name, api_call, api_arguments, environment_requirements, example_code, performance, and description.}. We provide more information in the Appendix. These fields were chose to generalize beyond the API calls within ML domain, to other domains, including RESTful API calls.” Here, Patil shows that the json object is the structured data object.
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Gnanasambandam and incorporate into the teachings of Patil because both references teach using an event progression model to process a natural language query to create a simulated event risk data object, then generate a structured data object by translating the natural language query into the structured data object using the large language model.
One of ordinary skill in the art would be motivated to do so because performing “Retriever-Aware training For training with retriever, the instruction-tuned dataset, also has an additional "Use this API documentation for reference: " appended to the user prompt. Through this, we aim to teach the LLM to parse the second half of the question to answer the first half. We demonstrate that this a) makes the LLM adapt to test-time changes in API documentation, and b) improves performance from in-context learning, and finally c) show that it reduces hallucination error,” (see Patil in page 5, paragraph 2).
Claim 11:
Regarding claim 11, Gnanasambandam teaches “A computing system comprising memory and one or more processors communicatively coupled to the memory, the one or more processors configured to:..”
See Gnanasambandam in [0004] describe “In some embodiments, a system includes a memory storing instructions and a processor communicatively coupled with the memory. The processor may execute the instructions to perform one or more of the operations described above.” Here, Gnanasambandam shows using a memory and processor to run instructions to perform operations of their system.
Regarding claim 11, it comprises of similar additional limitations as independent claim 1, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claim 12:
Regarding claim 12, Gnanasambandam in view of Patil, teach the limitations of claim 11.
Regarding claim 12, it comprises of similar additional limitations as corresponding claim 2, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claim 20:
Regarding claim 20, Gnanasambandam teaches “20. One or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors, cause the one or more processors to:…”
See Gnanasambandam in [0005] describe “In some embodiments, a tangible, non-transitory computer-readable medium stores instructions.” Here, Gnanasambandam describes using a non-transitory computer-readable medium.
Regarding claim 20, it comprises of similar additional limitations as independent claim 1, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claims 3 and 13 are rejected under 35 U.S.C. 103 as unpatentable over Gnanasambandam in view of Patil, further in view of Duarte J., et al., in “FAIR AI models in high energy physics. Machine Learning: Science and Technology,” published on December 29, 2023, available at: https://iopscience.iop.org/article/10.1088/2632-2153/ad12e3/pdf , (hereafter, Duarte).
Claim 3:
Regarding claim 3, Gnanasambandam in view of Patil, teach the limitations of claim 1. However, Gnanasambandam in view of Patil, did not teach “3. The computer-implemented method of claim 1, wherein the structured data object comprises a set of computer instructions for interacting with the event progression model.”
In an analogous art, Duarte teaches “3. The computer-implemented method of claim 1, wherein the structured data object comprises a set of computer instructions for interacting with the event progression model.”
See Duarte in page 4, describe “The questions that the repository asks the user upon project creation can be found and modified in the file cookiecutter.json. The Makefile contains commands that allow the user to do various things with their project, such as downloading the data, setting up the test environment, converting the dataset, and training and evaluating the model. It also contains global variables obtained from cookiecutter.json. This procedure makes it explicit that the analysis operations are a directed acyclic graph (DAG).” Here, Duarte mentions that the json file shows a dataset that is a structured data object. Examiner construes structured data object to be any form of information such as a dataset, or a file that contains information of variables or records.
Further, see Duarte in page 8, section 2.4.2.2. Deployment to DLHub describe “We have made the trained ML model accessible [62] and reusable for inference by making it publicly available via DLHub [43, 63]. DLHub provides a custom software development kit (SDK) called dlhub_sdk that allows users to package and preserve a trained model with necessary dependencies, including packages with specific versions, custom modules, and serialized data and model files. Once a model has been published, its dedicated API can be used to run remote inference tasks using funcX, a fire-and-forget remote function execution that elastically deploys workers and containers across nodes in clouds, clusters, and supercomputers [64]. The process of making a model available is simplified with a notebook template made available by DLHub developers. This notebook requires the user to implement the inference code as a function that is executed during model calls, and to declare model-specific dependencies and associate metadata. The notebook template is accompanied with a document template with necessary information about the model… The published model includes a DOI, list of authors, point of contact, relevant information about input and output data type and shape, and instructions to run the ML model with a sample test set.” Here, Duarte shows that the notebook template which is a part of DLHub software that contains various data files (including structured data object mentioned from page 4) and other packages, also contain information about instructions to run a machine learning model, and interact with the model.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the references of Gnanasambandam and Patil with the teachings of Duarte by using the teachings of Gnanasambandam and Patil, with Duarte’s teaching of a structured data object has a set of computer instructions for interacting with the event progression model.
One of ordinary skill in the art would be motivated to do so because by integrating Duarte’s framework into the methods of Gnanasambandam and Patil, one with ordinary skill in the art would achieve "ML models, ..., have been shown to dramatically improve the rejection of this background, while retaining high H → bb detection efficiency thus enabling the study of this decay mode," (see Duarte, on page 6, in section 2.4, paragraph 1), and a method that“ create an Apptainer container [52] to improve the model’s portability across platforms, and evaluate the model’s inference performance within the container,” (see Duarte in page 8, section 3. Results, part 3.1).
Claim 13:
Regarding claim 13, it comprises of similar additional limitations as corresponding claim 3, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claims 4 and 14 are rejected under 35 U.S.C. 103 as unpatentable over Gnanasambandam in view of Patil, further in view of Wu, Y. et al., in “B2: Bridging code and interactive visualization in computational notebooks,” published on October 20-23, 2020, for a conference, available at https://dl.acm.org/doi/pdf/10.1145/3379337.3415851 , (hereafter, Wu).
Claim 4:
Regarding claim 4, Gnanasambandam, in view of Patil, teach the limitations of claim 1. However, Gnanasambandam in view of Patil, did not teach “4. The computer-implemented method of claim 1, wherein generating the structured data object comprises: generating the structured data object by executing a read-evaluate-print (REPL) loop based on the natural language query.”
In an analogous field, Wu teaches “4. The computer-implemented method of claim 1, wherein generating the structured data object comprises: generating the structured data object by executing a read-evaluate-print (REPL) loop based on the natural language query.”
See Wu in page 155, in section Related work, subsection Interactions Generating Code describe “in particular, GUESS offers an environment where interactions with graph visualizations can be captured in Python based REPL (read-evaluate-print loop), and textual commands manipulate the visual output. DEVise identifies that interactive visualizations can be modeled as SQL expressions, and that multiple views can be coordinated by analyzing their schemas—an approach analogous to B2’s automatic synthesis ...” Here, Wu shows using a system called GUESS that identifies user interactions inside a python based read-evaluate-print loop or REPL loop, and allows users to type commands into the system, where commands relate to queries.
Further, see Wu in page 159, in section Visual Interactions to Code, first three paragraphs, describe “The data in the current selection is accessed through the get_filtered_data API, which returns a standard dataframe object. Code. The analyst also wishes to access the code that derives the data of the state histogram to (1) run the code on a different dataset with the same schema, (2) share or record the code so the result can be reproduced directly, or compared with other analysis. This code can be accessed through the get_code API, as well as the Copy Code to Clipboard dashboard button 11 .
Predicates. The analyst realizes that they also want to access selections that occurred previously. The all_selections API returns the full history of selections as a list of B2 objects, and a corresponding API returns the current_selection. These B2 objects can be reused using additional API calls that give access to either a data frame or code representation.” Here, Wu describes that using a standard dataframe object relates to using a structured data object. Later, Wu shows that this code can be accessed through API calls to retrieve the data frame.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the references of Gnanasambandam and Patil with the teachings of Wu by using the teachings of Gnanasambandam and Patil, with Wu’s teaching of running a read-evaluate-print (REPL) loop based on the natural language query.
One of ordinary skill in the art would be motivated to do so because by integrating Wu’s framework into the methods of Gnanasambandam and Patil, one with ordinary skill in the art would achieve a method of using “analysts… can use code comments to document meaningful interactive discoveries… qualitative comments indicate that B2 helps facilitate the exploratory data analysis process,” (See Wu in page 153, last two paragraphs of Introduction).
Claim 14:
Regarding claim 14, it comprises of similar additional limitations as corresponding claim 4, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claims 5 and 15 are rejected under 35 U.S.C. 103 as unpatentable over Gnanasambandam in view of Patil, further in view of Arora N. et al., in “A Bolasso based consistent feature selection enabled random forest classification algorithm: An application to credit risk assessment,” published on January 1st, 2020, available at https://www.sciencedirect.com/science/article/pii/S1568494619307173 , (hereafter, Arora).
Claim 5:
Regarding claim 5, Gnanasambandam, in view of Patil, teach the limitations of claim 1.
However, Gnanasambandam in view of Patil, did not teach “5. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: removing one or more features from the entity feature dataset based on the structured data; and generating an event risk score for the entity identifier based on a sequence of removal of the one or more features from the entity feature dataset.”
In an analogous art, Arora teaches “5. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: removing one or more features from the entity feature dataset based on the structured data object;”
See Arora in page 1, Abstract note “Credit risk assessment has been a crucial issue as it forecasts whether an individual will default on loan or not. Classifying an applicant as good or bad debtor helps lender to make a wise decision. The modern data mining and machine learning techniques have been found to be very useful and accurate in credit risk predictive capability and correct decision making. Classification is one of the most widely used techniques in machine learning. To increase prediction accuracy of standalone classifiers while keeping overall cost to a minimum, feature selection techniques have been utilized, as feature selection removes redundant and irrelevant attributes from dataset...The consistent feature selection is defined as robustness of selected features with respect to changes in dataset Bolasso generated shortlisted features are then applied to various classification algorithms like Random Forest (RF), Support Vector Machine (SVM), Naïve Bayes (NB) and K-Nearest Neighbors (K-NN) to test its predictive accuracy. It is observed that Bolasso enabled Random Forest algorithm (BS-RF) provides best results for credit risk evaluation.” Arora here mentions using a method called feature selection that removes attributes or features from the data, and this information is used to provide results for credit risk or viewed as calculating a risk score.
Further, see Arora in page 1, Introduction describe “Credit Risk is defined as a probability that a borrower will fail to repay the borrowed amount. The decision of granting or rejecting a loan is very critical and is based on applicant’s personal information, credit history, living status, loyalty etc.” Here, Arora mentions that predicting a credit risk, in this case as a form of probability score, is based on a person’s data such as credit history, personal information, etc. Examiner construes structured data to be any form of information such as a dataset, in this case a person’s data of personal information or credit history. The structured data object is construed here to include a person’s data.
See also Arora on page 5, section 4. Dataset description, 4.1. Lending Club mention “Lending club dataset [64] consists of 42,538 loan records having 143 attributes issued by lending club between years 2007–2011. Some of the important features describing loan are shown in Table 1.” Arora here mentions using a method called feature selection that removes attributes or features from the data, and this information is used based on loan records. The structured data here is the loan record.
Further, Arora teaches “and generating an event risk score for the entity identifier based on a sequence of removal of the one or more features from the entity feature dataset.”
See Arora in page 2, Introduction, mention “This paper proposes improved version of embedded based Lasso approach, namely Bolasso to shortlist consistent and relevant features. As Credit risk assessment is a difficult learning problem, so selection of appropriate set of features is critical for success of learning process and therefore is a vital issue. An important property of feature selection strategy is selection consistency (stability) i.e. to find “true” set of features [9]. The feature stability matters especially when feature selection is applied for knowledge discovery. In credit risk assessment, a feature selection algorithm may select different subsets of features when there are slight variations in training data, although most of these subsets may result in good prediction accuracy, such changes (instability) in selected features may reduce the confidence of lenders in investigating reliable risk factors.” Here, Arora explicitly describes that credit risk calculation for a person (i.e. relates to event risk score for the entity identifier) is based on feature selections (which includes feature removal as mentioned from page 1, abstract,) from a dataset.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the references of Gnanasambandam and Patil with the teachings of Arora by using the teachings of Gnanasambandam and Patil, with Arora’s teaching of generating an event risk score for the entity identifier based on a sequence of removal of the one or more features from the entity feature dataset.
One of ordinary skill in the art would be motivated to do so because by integrating Arora’s framework into the methods of Gnanasambandam and Patil, one with ordinary skill in the art would achieve “The conclusion drawn was keeping less features would help to reduce workload of credit evaluators and at the same time, increases predictive efficiency,” (see Arora in page 3, section 2.2 Feature selection based study in credit risk assessment).
Claim 15:
Regarding claim 15, it comprises of similar additional limitations as corresponding claim 5, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claims 6 and 16 are rejected under 35 U.S.C. 103 as unpatentable over Gnanasambandam in view of Patil, further in view of Zhan J., et al., in “Explainable artificial intelligence: Counterfactual explanations for risk-based decision-making in construction,” published on January 23, 2024, available at: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10413227 , (hereafter, Zhan).
Claim 6:
Regarding claim 6, Gnanasambandam in view of Patil, teach the limitations of claim 1. However, Gnanasambandam in view of Patil, did not teach “6. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: modifying a particular feature from the entity feature dataset in response to a determination, based on the structured data object, that a modification to the particular feature results in a greatest change to the event risk data object as compared to one or more other features from the entity feature dataset.”
In an analogous art, Zhan teaches “6. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: modifying a particular feature from the entity feature dataset in response to a determination, based on the structured data object, that a modification to the particular feature results in a greatest change to the event risk data object as compared to one or more other features from the entity feature dataset.”
See Zhan in page 10681, section 2) Feature Necessity Evaluation describe “It is crucial to determine whether a particular factor is indispensable for providing an explanation. In general, if altering one specific variable while holding other variables constant changes the predicted result, then that variable is necessary for the original risk prediction during tunnel construction. We also use SHAP to qualify the interpretation of ‘‘important” features [13]. The results of feature attribution and rankings are presented in Fig. 15(a). In addition, we utilize (17) to examine the necessity of features by fixing other features’ values and changing the selected features to check if the feature is necessary for the output within a specific range [see Fig. 15(b) and (c)]. Fig. 15 shows that feature importance rankings may generally align with each feature's necessity; they can also deviate from this trend, as we saw with the SHAP.” Here, Zhan shows that if modifying a variable caused a change in the mean SHAP value (measure relates to greatest change), or cause a change in the result for predicting risk during tunnel construction. Zhan identifies which factor or variable is indispensable or is necessary (i.e. cause the greatest change) for an original risk prediction (i.e. to an event risk data object), then that variable is compared among other variables. See Figure 15 for the comparison of variables for details.
Further, see Zhan in page 10669 item 3 in paragraph 2 mention “Thus, CF provide the ability to take action to cause a specific outcome.” Also, see Zhan in page 10669, section III. Research approach, mention “Our research focuses on the risk of ground settlement as they are identified as a critical contributor to the safety of tunnel construction. To achieve this goal, we combine a DiCE and DNN model to generate CFE to determine the importance, necessity, and sufficiency of factors tunnel-induced by ground settlement prediction and assessment, which help improve risk-based decision-making during construction. The workflow of our approach is presented in Fig. 1 and described below in three steps. 1) Construction of Kernel Principal Components Analysis based DNN model (KPCA-DNN): The input of the constructed KPCA-DNN model is the shield tunneling parameters (i.e., operational parameters and geological parameters). The output of KPCA-DNN is the value of tunnel-induced ground settlement during tunnel construction. 2) Generation of CFE: We use DiCE to generate CF within a desired range by adjusting the shield tunneling parameters. Four properties, diversity, proximity sparsity, and user constraints, are added to the generation of CF, which aim to provide diverse and practical CF with minimum adjustment from the original query instance. 3) Analyze risk prediction and assessment factors’ importance, necessity, and sufficiency: Here, based on the generated CF obtained in Step 2, we determine the factors’ importance, necessity, and sufficiency on tunnel-induced ground settlement prediction and assessment by counting valid CF…As a result, the DNN has been widely applied and used for risk prediction in tunnel construction due to its ability to capture abstract concepts and patterns in the data and handle nonlinear relationships between input and output variables.” Here, Zhan shows that the modification of a variable or feature is determined from the output of a risk based decision model. When Zhan mentions the ‘factors’ importance, necessity, and sufficiency on tunnel-induced ground settlement prediction and assessment by counting valid CF’, Zhan shows that based on the CF or counterfactual record (i.e. viewed as structured data object), this modifies a feature variable that causes the greatest change to a risk prediction (i.e. to the event risk data object).
PNG
media_image1.png
892
1065
media_image1.png
Greyscale
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the references of Gnanasambandam and Patil with the teachings of Zhan by using the teachings of Gnanasambandam and Patil, with Zhan’s teaching of modifying a particular feature from the entity feature dataset in response to a determination, based on the structured data object, that a modification to the particular feature results in a greatest change to the event risk data object.
One of ordinary skill in the art would be motivated to do so because by integrating Zhan’s framework into the methods of Gnanasambandam and Patil, one with ordinary skill in the art would achieve a method “by comparing KPCA-DNN and DNN, the KPCA process improves prediction robustness. In sum, our KPCA-DNN model achieves higher accuracy and better fitting,” (see Zhan in page 10672, section B, Results of KPCA-DNN-Based Risk Prediction).
Claim 16:
Regarding claim 16, it comprises of similar additional limitations as corresponding claim 6, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claims 7 and 17 are rejected under 35 U.S.C. 103 as unpatentable over Gnanasambandam in view of Patil, further in view of Karvanen, J. et al., in “Simulating counterfactuals,” published on March 26, 2024, as version 3, available at https://arxiv.org/html/2306.15328v3 , (hereafter, Karvanen).
Claim 7:
Regarding claim 7, Gnanasambandam in view of Patil, teach the limitations of claim 1. However, Gnanasambandam in view of Patil, did not teach “7. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: generating a simulated entity feature dataset by adding a simulated event feature to the entity feature dataset based on the structured data object; and generating an event risk score for the entity identifier based on the simulated entity feature dataset.”
In an analogous art, Karvanen teaches “7. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: generating a simulated entity feature dataset by adding a simulated event feature to the entity feature dataset based on the structured data object;”
See Karvanen in page 25, section 6. Discussion describe “We presented an algorithm for simulating data from a specified counterfactual distribution (Algorithm 4). The algorithm receives a known SCM and a counterfactual of interest as inputs and simulates independent but non-unique observations from the counterfactual distribution. The algorithm is intended to be a tool for simulation studies on counterfactual inference and fairness evaluation.” Here, Karvanen shows the concept of ‘simulating data’ shows a simulated entity feature dataset. By using ‘a counterfactual of interest as inputs and simulates independent but non-unique observations’, Karvanen shows generating a simulated observations (i.e. simulated entity feature dataset), by adding simulated observations or data. Counterfactual distribution is construed to also mean simulated.
Further, see Karvanen in page 5, section 2, Notation and Definitions, describe “Data can be simulated from an SCM by generating the values 𝐮 of the background variables 𝐔 from the distribution p(𝐮) and then applying the functions 𝐅 in the topological order to obtain values 𝐯 of 𝐕. For a data matrix 𝐃 with named columns (variables),” Here, Karvanen explicitly describes data such as background variables can be simulated, and relates to adding a simulated event feature to the simulated dataset.
Also, see Karvanen in pages 14-15, section 3.3 Algorithms for Counterfactual Inference and Fairness mention “Next, we present high-level algorithms that use Algorithm 3 to simulate data from a multivariate conditional distribution… In practice, the functions of variables 𝐗 in the SCM are replaced by constant-valued functions. On line 4, counterfactual observations are simulated from the intervened SCM with the background variables simulated on line 2.” Here, Karvanen shows adding simulated counterfactual observations in an algorithm. See Algorithm 4 for details.
PNG
media_image2.png
314
1056
media_image2.png
Greyscale
Further, Karvanen teaches “and generating an event risk score for the entity identifier based on the simulated entity feature dataset”
See Karvanen in page 22, section 5 Application to Fairness in Credit-Scoring describe “In this section, we show how Algorithm 5 can be used in the fairness analysis of a synthetic scenario where the details of the prediction model are unknown and real data are not available. We consider credit-scoring where the decision to grant a bank loan to a consumer depends on personal data and the amount of the loan. The credit decision is made by an automated prediction model that can be accessed via an application programming interface (API) but is otherwise unknown to the fairness evaluator. The outcome of the prediction model is the default risk.” Here, Karvanen shows that using the simulated data and simulated variables described, a default risk or event risk score is created for a person, which is the entity identifier. See page 37 in Online appendix 4 for details.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the references of Gnanasambandam and Patil with the teachings of Karvanen by using the teachings of Gnanasambandam and Patil, with Karvanen’s teaching of generating an event risk score for the entity identifier based on the simulated entity feature dataset.
One of ordinary skill in the art would be motivated to do so because by integrating Karvanen’s framework into the methods of Gnanasambandam and Patil, one with ordinary skill in the art would achieve a method with an “algorithm 3 remains effective for an SCM with a high total number of variables, as long as the number of conditioning variables J remains moderate,” (see Karvanen in page 20, section 3.4).
Claim 17:
Regarding claim 17, it comprises of similar additional limitations as corresponding claim 7, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claims 8 and 18 are rejected under 35 U.S.C. 103 as unpatentable over Gnanasambandam in view of Patil, and further in view of Aldughayfiq, B., et al., in “Layer-Weighted Attention and Ascending Feature Selection: An Approach for Seriousness Level Prediction Using the FDA Adverse Event Reporting System,” published on April 13, 2024, available at https://www.mdpi.com/2076-3417/14/8/3280, (hereafter, Aldughayfiq).
Claim 8:
Regarding claim 8, Gnanasambandam in view of Patil, teach the limitations of claim 1. However, Gnanasambandam in view of Patil, did not teach “8. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: determining importance of respective features of the entity feature dataset based on the structured data object; generating the one or more attention values based on the importance of respective features; generating a weighted entity feature dataset by weighting the entity feature dataset based on the one or more attention values; and generating, using the event progression model, the simulated event risk data object based on the weighted entity feature dataset.”
In an analogous art, Aldughayfiq teaches “8. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: determining importance of respective features of the entity feature dataset based on the structured data object;”
See Aldughayfiq describe in page 11, section, 4.3. Feature Selection “Feature selection is paramount in refining the predictive power of a model by narrowing it down to the most salient features while discarding the redundant or less informative ones. In this study, we relied on the robust capabilities of the Select_K_Best method." Here, Aldughayfiq shows using a method for feature selection to identify the most salient or most important features of a dataset.
See page 1, in Introduction, where Aldughayfiq mentioned "Traditionally, such problems have been tackled through the selection of a subset of features that are deemed to be most relevant based on prior knowledge or statistical analysis." Here, Aldughayfiq mentions selection of a subset of features that is most relevant to relate to determine the importance of features.
Further, See Aldughayfiq in page 13, section 6 Hybrid Ascending Feature Selection with Layer-Static-Weighted Attention Algorithm (HAFS-LWA) mention “Figure 10 displays the weight distributions for the deep learning model’s features at three distinct attention weights (α = 0.1, 0.5, and 0.9), with each row of histograms corresponding to a different attention weight setting that balances the focus between NLP and demographic features. At α = 0.1 (Figure 10a), the attention weight favors demographic features (90%) over NLP features (10%)," Here, Aldughayfiq shows identifying importance of variables or features of an entity dataset using a data object (of dataset) based on attention weights. See Figure 10, page 16 where Aldughayfiq shows feature relevance by weight values.
PNG
media_image3.png
847
617
media_image3.png
Greyscale
Further, Aldughayfiq teaches “generating the one or more attention values based on the importance of respective features;”
See Aldughayfiq in page 13, section 6, describe “One branch goes to a “demographic features” block, which contains some additional features that are specific to the task. The other branch goes to an “attention weight” block. The attention weight block calculates a weight for each of the 64 features in the input. This weight represents how important that feature is for the task. The weights are then used to combine the features from the “NLP” and “demographic features” blocks into a single vector of 64 features.” Here, Aldughayfiq shows that ‘The attention weight block calculates a weight for each of the 64 features in the input,’ which relates to creating attention values based on importance of each of the features.
Further, see Aldughayfiq mention in abstract, page 1 “Simultaneously, we employed a layer-static-weighted attention technique, which dynamically adjusts the model’s focus between natural language processing (NLP) and demographic features. This technique achieved its best performance at a balanced weight of 50%, yielding an average test accuracy of 74.56% and CV ROC score of 0.83 when 4000 features were included, indicating a compelling advantage to include a larger volume of meaningful features. By integrating these methodologies, we constructed a robust model capable of effectively predicting seriousness levels, offering significant potential for improving pharmacovigilance and enhancing drug safety monitoring.” Here, Aldughayfiq shows that predicting seriousness levels relate to simulated event risk data object, since seriousness levels are predicted or simulated based on a model. This prediction is also based on weighted attention of weights on features.
Also, see Aldughayfiq in page 4, section 2. Related works, describe “Another research methodology regarding the measurement of the adverse effects of the COVID-19 vaccine is presented in [21]. The primary objective of this work was to analyze vaccine adverse event data using ontologies and ML. To make it simple for users to access and obtain these data, the researchers also created an intuitive user interface.” Here, Aldughayfiq shows that by using a model and event data from vaccine records (i.e. structured data object), a simulated risk level was determined from this information.
Further, Aldughayfiq teaches “generating a weighted entity feature dataset by weighting the entity feature dataset based on the one or more attention values;”
See Aldughayfiq in page 13, section 6, describe “One branch goes to a “demographic features” block, which contains some additional features that are specific to the task. The other branch goes to an “attention weight” block. The attention weight block calculates a weight for each of the 64 features in the input. This weight represents how important that feature is for the task. The weights are then used to combine the features from the “NLP” and “demographic features” blocks into a single vector of 64 features.” Here, Aldughayfiq shows that ‘The attention weight block calculates a weight for each of the 64 features in the input,’ which relates to creating attention values based on importance of each of the features.
Also, see Aldughayfiq describe in pages 19-20 “The layer-static-weighted attention method demonstrated the intricate balance required between NLP-driven and demographic features. Intriguingly, the results peaked when both feature types were given equal emphasis, registering the highest average test accuracy of 70.286% and CV ROC score of 74.49% at a 50% weightage. In light of the empirical data, a clear takeaway emerges: While the sheer volume of features plays a pivotal role in enhancing performance, the judicious weighting between different feature categories is equally crucial. This dual strategy ensures not only a granular understanding of the data but also a model that is attuned to the nuanced variances within.” Here, Aldughayfiq shows weighting the whole dataset (i.e. create weighted entity) using layer static weighted attention method using attention values.
Also, see Aldughayfiq mention in abstract, page 1 “Simultaneously, we employed a layer-static-weighted attention technique, which dynamically adjusts the model’s focus between natural language processing (NLP) and demographic features. This technique achieved its best performance at a balanced weight of 50%, yielding an average test accuracy of 74.56% and CV ROC score of 0.83 when 4000 features were included, indicating a compelling advantage to include a larger volume of meaningful features. By integrating these methodologies, we constructed a robust model capable of effectively predicting seriousness levels, offering significant potential for improving pharmacovigilance and enhancing drug safety monitoring.” Here, Aldughayfiq shows that predicting seriousness levels relate to simulated event risk data object, since seriousness levels are predicted or simulated based on a model. This prediction is also based on weighted attention of weights on features.
Further, see Aldughayfiq in page 4, section 2. Related works, describe “Another research methodology regarding the measurement of the adverse effects of the COVID-19 vaccine is presented in [21]. The primary objective of this work was to analyze vaccine adverse event data using ontologies and ML. To make it simple for users to access and obtain these data, the researchers also created an intuitive user interface.” Here, Aldughayfiq shows that by using a model and event data from vaccine records (i.e. structured data object), a simulated risk level was determined from this information.
Further, Aldughayfiq teaches “and generating, using the event progression model, the simulated event risk data object based on the weighted entity feature dataset”
See Aldughayfiq in page 13, section 6, describe “One branch goes to a “demographic features” block, which contains some additional features that are specific to the task. The other branch goes to an “attention weight” block. The attention weight block calculates a weight for each of the 64 features in the input. This weight represents how important that feature is for the task. The weights are then used to combine the features from the “NLP” and “demographic features” blocks into a single vector of 64 features.” Here, Aldughayfiq shows that ‘The attention weight block calculates a weight for each of the 64 features in the input,’ which relates to creating attention values based on importance of each of the features.
Further, see Aldughayfiq in page 1, abstract describe “we employed a layer-static-weighted attention technique, which dynamically adjusts the model’s focus between natural language processing (NLP) and demographic features. This technique achieved its best performance at a balanced weight of 50%, yielding an average test accuracy of 74.56% and CV ROC score of 0.83 when 4000 features were included, indicating a compelling advantage to include a larger volume of meaningful features. By integrating these methodologies, we constructed a robust model capable of effectively predicting seriousness levels, offering significant potential for improving pharmacovigilance and enhancing drug safety monitoring. The results underscore the value of NLP and demographic data in predicting drug event seriousness” Here, Aldughayfiq shows for each feature of the data in the model (i.e. event progression model), the model generates drug safety monitoring in predicting seriousness levels (i.e. predicting simulated event risk data object) based on weighted dataset described in page 13, section 6 of weighted attention block for each features in the input data (i.e. weighted entity feature dataset).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the references of Gnanasambandam and Patil with the teachings of Aldughayfiq by using the teachings of Gnanasambandam and Patil, with Aldughayfiq’s teaching of generating the one or more attention values based on the importance of respective features for a weighted entity feature dataset .
One of ordinary skill in the art would be motivated to do so because by integrating Aldughayfiq’s framework into the methods of Gnanasambandam and Patil, one with ordinary skill in the art would achieve the goal of “by integrating these methodologies, we constructed a robust model capable of effectively predicting seriousness levels, offering significant potential for improving pharmacovigilance and enhancing drug safety monitoring,” (see Aldughayfiq, in page 1, abstract).
Claim 18:
Regarding claim 18, it comprises of similar additional limitations as corresponding claim 8, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claims 9 and 19 are rejected under 35 U.S.C. 103 as unpatentable over Gnanasambandam in view of Patil, and further in view of Abu El Ata, N. et al., published on March 10, 2022, US PG Pub. No. US20220076841A1, (hereafter, Abu).
Claim 9:
Regarding claim 9, Gnanasambandam in view of Patil, teach the limitations of claim 1. However, Gnanasambandam in view of Patil, did not teach “ 9. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: generating an event risk score for respective features of the entity feature dataset based on the structured data object;” or “and generating the simulated event risk data object based on the event risk score”,
In an analogous art, Abu teaches “ 9. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: generating an event risk score for respective features of the entity feature dataset based on the structured data object;”
See Abu in [0058] describe “Health attributes 106 of the patient 105 may then be obtained (210). The health attributes may include a range of data, test results, evaluation and analysis of the patient 105, including genetic traits 108 a, a blood lipid profile 108 b, medical imaging reports 108 c, medical records 108 d detailing the patient's medical history, and the patient's environment 108 e (e.g., climate, location, and related external factors that may affect the patient's health).” Here, Abu shows the data object includes a medical record of a patient, including a patient’s medical profile such as lipid profile, imaging, medical history, etc.
Also, see Abu mention in [0059-0060] “Once the patient model 120 is complete, it may be modeled (simulated) under plural sets of parameters. The plural sets of parameters may be previously generated as detailed with reference to the example embodiments described below (220), and each of the plural sets of parameters may indicate respective variables having a causal relation to the health metrics. For example, one set of parameters may represent exposure of the patient to a disease, a change in the patient's health attributes based on known trends within the population of the patient's environment, an injury suffered by the patient, and other changes relating to the patient or the patient's environment that may affect the patient's health. Further sets of parameters may represent a control scenario (i.e., a continuation of the patient's experience without introduction of health-affecting variables), as well as one or more interventions (e.g., scenarios in which a patient makes positive changes to his/her health attributes, such as quitting smoking, changing a diet or exercise routine, engaging in physical therapy, or taking a medication). The health metrics of the patient model 120 are then modeled under the plural sets of parameters to generate respective health metrics (e.g., resistance metric, HDD) (225). [0060] From an analysis of the plural sets of parameters and the resulting health metrics, occurrence probabilities for each of the sets of parameters can be determined (230). The occurrence probabilities may indicate predicted probabilities of the patient model transitioning from an initial state to each of a plurality of successive states, each of the successive states corresponding to a respective one of the plural sets of parameters. Some of those successive states may be identified to relate to an adverse outcome, such as a diseased state representing the patient being afflicted by a disease, or an outcome wherein the patient's health metrics decrease below a given threshold. Accordingly, one or more risks may also be identified, wherein the risks indicate a likelihood of the patient 105 (represented by the model 120) transitioning from the initial state (corresponding to the initial health attributes 106) to an adverse outcome such as a diseased state.” Here, Abu mentions the states (i.e. each of the features) may be identified to relate to an adverse outcome (i.e. event risk score) where adverse outcome is similar to a risk score (i.e. generating an event risk score for respective features of the entity feature dataset based on the structured data object) where structured object here shows the patient’s health metrics.
Later, see Abu in [0127-0128] describe “Example embodiments may utilize the knowledge database to continuously monitor patient models to cover the risk of both the knowns as well as the unknowns (e.g., risks) that are caused by the evolutionary nature of dynamic complexity. [0128] In phase five, risk monitoring 2825, the monitoring process is implemented. Using the database that contains all risk cases generated in phase three 2815 and enhanced with remedial plans in phase four 2820, the patient model may be put under surveillance using automation technologies.” Here, Abu El Ata mentions using a knowledge database (i.e. structured data object) that cover risk monitoring of cases and are connected to identifying adverse outcomes.
2)Further, Abu teaches “and generating the simulated event risk data object based on the event risk score”
See Abu El Ata in [0061] describe “Example embodiments, as described herein, provide for emulating a patient model 120 through a number of differing scenarios, where the results of such emulations can be analyzed to identify health-related concerns (e.g., susceptibility to a disease) and potential remedies for the patient. One aspect of this emulation, as described above, is to generate plural sets of parameters (220). In order to generate those sets of parameters, a set of input parameters may be permutated, by altering one or more values, to generate one or more additional scenarios for emulation. Such selection of differing parameters is described herein, and in particular with reference to FIG. 7. When selecting input parameters to detect an adverse outcome resulting from a patient model's dynamic complexity, a number of variables can be selected for permutation. For example, input parameters can be permutated to simulate the patient catching a given disease, a change in one or more of the patient's health attributes (e.g., a change in the patient's blood lipid profile, a change in the patient's lifestyle or environment), an acute adverse outcome such as an injury, and/or one or more interventions, such as treatment of physical therapy, a medication, or a positive lifestyle change (e.g., quitting smoking, beginning an exercise routing or diet).” Here, Abu shows generating emulations or simulated parameters that can detect adverse outcomes (i.e. simulated risk data object) based on the input parameters such as an acute adverse outcome risk factor (i.e. event risk score).
Note the examiner construes simulated risk data object to be any data record or information that contains information about risk or probability of developing an adverse event from specification in [0062] mention “In some embodiments, the term “event risk data object” refers to a data entity that describes a machine learning prediction for event risk for respective predefined events. In some embodiments, an event risk data object may indicate a predicted degree of risk for respective events. In some embodiments, an event risk data object may include one or more event risk scores for one or more potential events. An event risk score may provide a predicted risk or probability that a particular event will occur at a future instance in time.” Event risk data object is viewed as a numerical value or a prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the references of Gnanasambandam and Patil with the teachings of Abu by using the teachings of Gnanasambandam and Patil, with Abu’s teaching of generating an event risk score for respective features of the entity feature dataset based on the structured data object, and then use the event risk score to create the simulated event risk data object.
One of ordinary skill in the art would be motivated to do so because by integrating Abu’s framework into the methods of Gnanasambandam and Patil, one with ordinary skill in the art would achieve a “process 400 may also identify effective remedies/interventions by identifying a causal relation between those remedies/interventions and an improvement to the outcome of the patient model under the modeled scenarios,” (see Abu El Ata in [0070]).
Claim 19:
Regarding claim 19, it comprises of similar additional limitations as corresponding claim 9, and is rejected under the same rationale under 35 U.S.C. 103 as applied hereinabove.
Claim 10 is rejected under 35 U.S.C. 103 as unpatentable over Gnanasambandam in view of Patil, and further in view of Fornwalt, B. et al., in US PG Pub. No. US20230087969A1, published on March 23, 2023, (hereafter, Fornwalt).
Claim 10:
Regarding claim 10, Gnanasambandam in view of Patil, teach the limitations of claim 1.
However, Gnanasambandam in view of Patil, did not teach “10. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises: generating a modified entity feature dataset by modifying one or more features of the entity feature dataset based on the structured data object; and inputting the modified entity feature dataset to the event progression model to generate the simulated risk data object.”
In an analogous art, Fornwalt teaches “10. The computer-implemented method of claim 1, wherein generating the simulated event risk data object comprises:
generating a modified entity feature dataset by modifying one or more features of the entity feature dataset based on the structured data object; and inputting the modified entity feature dataset to the event progression model to generate the simulated risk data object”
See Fornwalt in [0071-0072] describe “At 104, the process 100 can determine a first risk score for the patient based on the health information using a trained predictor model. The trained predictor model can be a linear model such as linear logistic regression or a non-linear model such as random forest or XGBoost as described above. The predictor model can be trained to predict risk scores of all-cause mortality for a predetermined time period, such as one year, although it is appreciated that the model could be trained to predict all-cause mortality for other time periods six months, two years, three years, four years, five year, or other appropriate time periods or other appropriate clinical endpoints. The process 100 can provide at least a portion of the health information to the model and receive the first risk score from the model. The first risk score can represent a baseline score corresponding to an actual predicted mortality risk of the patient. The process 100 can then proceed to 106. [0072] At 106, the process 100 can determine a second risk score for the patient based on the health information and at least one artificially closed care gap included in the health information using the predictor model. The process 100 can artificially close appropriate care gaps by changing the value of each open care gap from 1=open/untreated to 0=closed/treated while keeping all other variables included in the health information unchanged. The process 100 may not close certain care gaps in patients who meet the exclusion criteria for that care gap. For example, a patient with bradycardia who could not be treated with EBBB would not have the EBBB care gap closed. The process 100 can then provide the health information, which has been modified to close any appropriate care gaps, to the model and receive the second risk score from the model. The second risk score can represent a simulated score corresponding to what the predicted mortality risk of the patient would be if all appropriate open care gaps are closed.” Here, Fornwalt mentions that the system takes a patient variable of a care gap, and change them from open to closed, this maps to claim 10’s limitation of modifying one or more features of the entity feature dataset and then inputting the modified dataset (with the modified variable feature) to create another simulated score.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the references of Gnanasambandam and Patil with the teachings of Fornwalt by using the teachings of Gnanasambandam and Patil, with Fornwalt’s teaching of modifying one or more features of the entity feature dataset based on the structured data object; and inputting the modified entity feature dataset to the event progression model to generate the simulated risk data object.
One of ordinary skill in the art would be motivated to do so because by integrating Fornwalt’s framework into the methods of Gnanasambandam and Patil, one with ordinary skill in the art would achieve a method that “showed marked differences in outcomes and different responses to medications among the 4 subgroups, which could help to define effective treatment strategies specific to each subgroup,” (See Fornwalt in [0066]), and “these nonlinear models were hypothesized to improve predictive accuracy by capturing more complex, non-linear relationships among input variables. The best performing model was selected for subsequent analysis of care gap closure effect estimation,” (see Fornwalt in [0042]).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WENWEI ZENG whose telephone number is (571)272-7111. The examiner can normally be reached Monday-Friday, 8am-5pm.
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, Usmaan Saeed can be reached at (571) 272-4046. 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.
/WenWei Zeng/Examiner, Art Unit 2146
/USMAAN SAEED/Supervisory Patent Examiner, Art Unit 2146
Prosecution Insights