Patent Issued for Systems, devices, and methods for parallelized data structure processing (USPTO 11853360): Massachusetts Mutual Life Insurance Company

2024 JAN 16 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- Massachusetts Mutual Life Insurance Company (Springfield, Massachusetts, United States) has been issued patent number 11853360, according to news reporting originating out of Alexandria, Virginia, by NewsRx editors.

The patent’s inventors are Merritt, Sears (Groton, MA, US).

This patent was filed on September 2, 2022 and was published online on December 26, 2023.

From the background information supplied by the inventors, news correspondents obtained the following quote: “In this disclosure, where a document, an act, and/or an item of knowledge is referred to and/or discussed, then such reference and/or discussion is not an admission that the document, the act, and/or the item of knowledge and/or any combination thereof was at a priority date, publicly available, known to a public, part of common general knowledge, and/or otherwise constitutes any prior art under any applicable statutory provisions; and/or is known to be relevant to any attempt to solve any problem with which this disclosure may be concerned with. Further, nothing is disclaimed.

“A server may serve a network page to a client. The network page may include a set of fields programmed to receive a plurality of inputs from the client, such as a plurality of alphanumeric strings or a plurality of binary values. The network page may be further programmed to submit the inputs to the server, such as when the fields are populated or when triggered via the client. For example, a webserver may serve a webpage to a smartphone, where the webpage is programmed to receive a set of user inputs from the smartphone, such as personal information, address, health information, and others, and upload the user inputs to the webserver.

“Upon receiving the inputs from the client, the server may create a user profile based on the inputs and provide an output to the client based on the user profile. However, since some of the inputs may contain incorrect or imprecise information, some of the inputs may need to be verified or validated, such as independently. Therefore, until such verification or validation, the user profile may be classified as not reliable. Such classification may also taint the output in a similar light.

“When some of the inputs are determined to contain incorrect or imprecise information and when some of such inputs are amended in the user profile with correct or precise information, then this amendment updates the profile. Consequently, the output may be also be updated to account for the amendment. However, if the output has already been used in various data operations, then the update to the output may entail a repetition of such data operations with the output, as updated. This repetition wastes time and resources, such as computational cycles, memory space, and network bandwidth, especially cumulatively. If the output has not yet been used in various data operations, then a delay in verification or validation of some of the inputs is impractical. Accordingly, there is a desire for a computing technology to address at least one of such challenges.”

Supplementing the background information on this patent, NewsRx reporters also obtained the inventors’ summary information for this patent: “This disclosure at least partially addresses at least one of the above challenges. However, this disclosure can prove useful to other technical areas. Therefore, at least some claims should not be construed as necessarily limited to addressing any of the above challenges.

“In one embodiment, a computer-implemented method comprising: performing, by a server, a hyperparameter selection based on a parallelized grid search algorithm configured to determine a number of decision trees, a depth of decision trees, and an amount of variables used during tree node splitting; randomly selecting, by the server, a first subset from a data set, wherein the data set contains a set of records with a plurality of fields, wherein each of the fields includes a value; constructing, by the server, a learned mortality model for each of the number of decision trees, the depth of decision trees, and the number of variables based on the values of the first subset; identifying, by the server, a second subset from the data set, wherein the second subset is exclusive of the first subset; evaluating, by the server, the learned mortality model based on the second subset via at least one of a concordance or a time-varying area under curve statistic such that a first mortality score is determined; presenting, by the server, a user interface for display on a client computer based on the first mortality score satisfying a predetermined threshold, wherein the user interface is configured to receive a user input from the client computer; receiving, by the server, the input from the client computer, wherein the input includes a user profile with an attribute; submitting, by the server, the input into the learned mortality model such that the model outputs a second mortality score based on the user profile with the attribute; and updating, by the server, the user interface for display on the client computer with the second mortality score.

“In one embodiment, a memory storing a set of instructions for execution by a processor, wherein the set of instructions instructs the processor to: perform a hyperparameter selection based on a parallelized grid search algorithm configured to determine a number of decision trees, a depth of decision trees, and an amount of variables used during tree node splitting; randomly select a first subset from a data set, wherein the data set contains a set of records with a plurality of fields, wherein each of the fields includes a value; construct a learned mortality model for each of the number of decision trees, the depth of decision trees, and the number of variables based on the values of the first subset; identify a second subset from the data set, wherein the second subset is exclusive of the first subset; evaluate the learned mortality model based on the second subset via at least one of a concordance or a time-varying area under curve statistic such that a first mortality score is determined; present a user interface for display on a client computer based on the first mortality score satisfying a predetermined threshold, wherein the user interface is configured to receive a user input from the client computer; receive the input from the client computer, wherein the input includes a user profile with an attribute; submit the input into the learned mortality model such that the model outputs a second mortality score based on the user profile with the attribute; and update the user interface for display on the client computer with the second mortality score.

“Additional features and advantages of various embodiments are set forth in a detailed description which follows. Various objectives and other advantages of this disclosure are realized and attained by various structures particularly pointed out in various illustrative embodiments in the detailed description and claims hereof as well as in a set of appended drawings. Note that the detailed description is illustrative and explanatory and is intended to provide further explanation of this disclosure as claimed.”

The claims supplied by the inventors are:

“1. A method comprising: executing, by a server using a set of attributes of a user, a learned mortality model to predict a mortality score, the learned mortality model configured to: determine a number of decision trees, a depth of decision trees, and an amount of variables used during tree node splitting; select a first subset of the set of attributes, the first subset of the set of attributes corresponding to a hyper-parameter attribute selected by the learned mortality model; configure at least one decision tree based on the first subset of the set of attributes; and execute the learned mortality model using a second subset of the set of attributes to determine the mortality score; and transmitting, by the server, the mortality score to an electronic device.

“2. The method of claim 1, wherein the learned mortality model executes a hyper-parameter attribute selection protocol to determine the number of decision trees, the depth of decision trees, and the amount of variables used during tree node splitting.

“3. The method of claim 2, wherein the hyper-parameter attribute selection protocol uses a parallelized grid search algorithm.

“4. The method of claim 1, wherein the first subset of the set of attributes is randomly selected.

“5. The method of claim 1, wherein at least one of the first subset of the set of attributes or the second subset of the set of attributes are received by the server.

“6. The method of claim 1, wherein the learned mortality model determines the mortality score using at least one of a concordance and a time-varying area under curve statistic algorithm.

“7. The method of claim 1, wherein a number of attributes of the first subset of the set of attributes or the second subset of the set of attributes is predetermined.

“8. The method of claim 1, wherein the learned mortality model is trained via a combination of a cost regression and a random survival forest protocol.

“9. The method of claim 1, wherein the second subset of the set of attributes is exclusive of the first subset of the set of attributes.

“10. The method of claim 1, wherein the set of decision trees is arranged based on historical mortality data corresponding to historical user attributes and their respective mortality data.

“11. A computer system comprising: a computer readable medium having one or more instructions that when executed cause a processor to: execute, using a set of attributes of a user, a learned mortality model to predict a mortality score, the learned mortality model configured to: determine a number of decision trees, a depth of decision trees, and an amount of variables used during tree node splitting; select a first subset of the set of attributes, the first subset of the set of attributes corresponding to a hyper-parameter attribute selected by the learned mortality model; configure at least one decision tree based on the first subset of the set of attributes; and execute the learned mortality model using a second subset of the set of attributes to determine the mortality score; and transmit the mortality score to an electronic device.

“12. The computer system of claim 11, wherein the learned mortality model executes a hyper-parameter attribute selection protocol to determine the number of decision trees, the depth of decision trees, and the amount of variables used during tree node splitting.

“13. The computer system of claim 12, wherein the hyper-parameter attribute selection protocol uses a parallelized grid search algorithm.

“14. The computer system of claim 11, wherein the first subset of the set of attributes is randomly selected.

“15. The computer system of claim 11, wherein at least one of the first subset of the set of attributes or the second subset of the set of attributes are received by the server.

“16. The computer system of claim 11, wherein the learned mortality model determines the mortality score using at least one of a concordance and a time-varying area under curve statistic algorithm.

“17. The computer system of claim 11, wherein a number of attributes of the first subset of the set of attributes or the second subset of the set of attributes is predetermined.

“18. The computer system of claim 11, wherein the learned mortality model is trained via a combination of a cost regression and a random survival forest protocol.

“19. The computer system of claim 11, wherein the second subset of the set of attributes is exclusive of the first subset of the set of attributes.

“20. The computer system of claim 11, wherein the set of decision trees is arranged based on historical mortality data corresponding to historical user attributes and their respective mortality data.”

For the URL and additional information on this patent, see: Merritt, Sears. Systems, devices, and methods for parallelized data structure processing. U.S. Patent Number 11853360, filed September 2, 2022, and published online on December 26, 2023. Patent URL (for desktop use only): https://ppubs.uspto.gov/pubwebapp/external.html?q=(11853360)&db=USPAT&type=ids

(Our reports deliver fact-based news of research and discoveries from around the world.)